MCQ Question set 2 Flashcards
What is a key challenge in AI governance regarding automated decision-making?
A) Ensuring AI decisions remain explainable and accountable
B) Reducing regulatory oversight for AI deployment
C) Making AI models as complex as possible
D) Keeping AI decision-making confidential from stakeholders
A) Ensuring AI decisions remain explainable and accountable - AI decision-making must be transparent to avoid unintended consequences.
A financial AI model is compromised by an adversarial attack, causing it to misclassify fraudulent transactions as legitimate. What is the best governance response?
A) Implement adversarial training and enhance anomaly detection capabilities
B) Maintain AI-driven fraud detection since it reduces operational costs
C) Reduce AI transparency to prevent revealing security gaps
D) Expand AI fraud detection without addressing security concerns
A) Implement adversarial training and enhance anomaly detection capabilities - AI security must be strengthened against adversarial manipulation.
Which of the following best describes the purpose of AI accountability in governance frameworks?
A) Ensuring organizations take responsibility for AI-driven decisions and their consequences
B) Allowing AI to operate autonomously without legal or ethical constraints
C) Keeping AI decision-making confidential from regulatory bodies
D) Eliminating human oversight in AI decision-making
A) Ensuring organizations take responsibility for AI-driven decisions and their consequences - AI accountability ensures ethical and legal responsibility.
Under the EU AI Act, which entity bears primary responsibility for ensuring AI system compliance with transparency obligations?
A) The AI provider
B) The end user of the AI system
C) The national regulatory authority
D) The European Artificial Intelligence Board
B) The AI provider - Providers must ensure that AI systems comply with transparency and explainability requirements before deployment.
An AI-powered content recommendation engine increasingly promotes sensationalist and misleading articles because they generate high engagement. What governance strategy should the company adopt?
A) Introduce content credibility scoring and diversify recommendation criteria
B) Increase AI automation to maximize engagement further
C) Reduce transparency about recommendation algorithms
D) Allow market demand to dictate AI content recommendations
A) Introduce content credibility scoring and diversify recommendation criteria - AI content recommendations should prioritize accuracy and ethical distribution.
Which type of bias occurs when an AI system reflects disparities found in its training data?
A) Historical bias
B) Statistical bias
C) Computational bias
D) Speed bias
A) Historical bias - This type of bias occurs when AI models reinforce existing inequalities present in training datasets.
What is the purpose of differential privacy in AI systems?
A) To ensure AI models do not memorize individual data points
B) To increase AI efficiency by collecting more user data
C) To eliminate the need for human oversight in AI decisions
D) To make AI models completely transparent to the public
A) To ensure AI models do not memorize individual data points - Differential privacy protects individual user data while maintaining model accuracy.
What is the primary function of notified bodies in AI compliance?
A) Conducting third-party conformity assessments for high-risk AI systems
B) Deploying AI models in regulated industries
C) Developing AI regulations and standards
D) Serving as internal compliance officers for AI providers
A) Conducting third-party conformity assessments for high-risk AI systems - Notified bodies provide independent verification of AI compliance.
Which fairness mitigation strategy involves modifying training data to remove bias before model training?
A) Pre-processing bias mitigation
B) In-processing bias mitigation
C) Post-processing bias mitigation
D) Proxy bias elimination
B) Pre-processing bias mitigation - This approach ensures that AI models are trained on fair and representative data.
A social media AI recommendation system prioritizes controversial content because it drives engagement. What is the best governance action?
A) Implement content fairness measures to balance recommendations
B) Maintain current recommendations to maximize platform revenue
C) Reduce transparency in content ranking to prevent criticism
D) Expand AI-driven content recommendations without modification
A) Implement content fairness measures to balance recommendations - AI in content moderation should prioritize credibility over engagement.
A government agency deploys AI for public benefits eligibility assessments. Reports indicate that applicants from certain ethnic backgrounds face more rejections. What is the most responsible governance response?
A) Conduct an audit and introduce fairness constraints in AI decision-making
B) Maintain the current system as it improves application processing speed
C) Reduce transparency in eligibility assessments to prevent appeals
D) Expand AI decision-making without reviewing potential bias
A) Conduct an audit and introduce fairness constraints in AI decision-making - AI in public services must be fair and free from discrimination.
Which of the following is the BEST example of AI transparency in governance?
A) Providing users with explanations for AI-generated decisions
B) Keeping AI algorithms confidential to protect intellectual property
C) Removing human intervention to make AI fully autonomous
D) Reducing regulatory oversight to improve AI efficiency
A) Providing users with explanations for AI-generated decisions - AI transparency ensures accountability and trust.
Which AI governance principle ensures that AI models are regularly evaluated and updated to prevent unintended consequences?
A) Model drift
B) Algorithmic accountability
C) Data minimization
D) AI scalability
B) Algorithmic accountability - AI systems must be continuously monitored to ensure they remain fair, safe, and effective.
A government agency wants to use AI to predict unemployment trends. To ensure fairness, what should be the top priority?
A) Removing all demographic information from the training data
B) Ensuring diverse and representative data collection
C) Using only historical employment data without modifications
D) Relying solely on private sector AI tools for forecasting
B) Ensuring diverse and representative data collection - AI predictive models must be built on unbiased, representative datasets to ensure fairness.
Which governance measure ensures AI models comply with ethical and legal standards?
A) Conducting regular audits and risk assessments
B) Reducing transparency in AI decision-making
C) Eliminating human oversight to increase efficiency
D) Allowing AI to operate autonomously without regulation
A) Conducting regular audits and risk assessments - Regular reviews ensure AI compliance with ethical and regulatory standards.
What is the main goal of AI fairness in governance frameworks?
A) To maximize AI efficiency at the cost of inclusivity
B) To prevent AI systems from producing discriminatory outcomes
C) To allow AI to operate without human intervention
D) To increase AI complexity for better predictive accuracy
B) To prevent AI systems from producing discriminatory outcomes - AI fairness ensures that automated decisions do not disadvantage any group.
Which of the following regulatory approaches is considered the MOST stringent for AI compliance?
A) Self-regulation by industry
B) Voluntary compliance frameworks
C) Government-mandated risk assessments and audits
D) No formal regulations
C) Government-mandated risk assessments and audits - This approach ensures compliance, accountability, and minimizes AI-related risks.
A university implements an AI grading system but finds that it disproportionately lowers scores for essays with creative arguments. What should the institution do?
A) Revise the AI model to recognize and fairly evaluate creative writing
B) Maintain the current grading model for consistency
C) Reduce transparency in grading criteria to prevent student appeals
D) Automate all grading decisions without human review
A) Revise the AI model to recognize and fairly evaluate creative writing - AI grading should support diverse learning styles and academic fairness.
Which security measure can help protect AI models from adversarial attacks?
A) Implementing adversarial training techniques
B) Reducing transparency in AI decision-making
C) Eliminating all regulatory constraints on AI systems
D) Allowing AI to operate without encryption
A) Implementing adversarial training techniques - These techniques help AI models recognize and resist adversarial manipulations.
A healthcare provider uses an AI-based triage system to prioritize patient treatment. A study finds that the model systematically assigns lower priority scores to older patients, affecting their access to timely medical care. What governance measure should be prioritized?
A) Review AI decision criteria and retrain the model to ensure fair patient prioritization
B) Maintain AI-driven triage decisions since they improve efficiency
C) Reduce AI transparency to prevent patient complaints
D) Expand AI-driven healthcare automation without modifications
B) Review AI decision criteria and retrain the model to ensure fair patient prioritization - AI in healthcare should not discriminate based on age or demographic factors.
An autonomous drone delivery service faces criticism after repeatedly failing to service lower-income areas due to perceived safety risks. How should the company respond?
A) Adjust the AI model to ensure equitable service distribution
B) Continue AI operations as optimizing safety is the primary concern
C) Remove human oversight to allow AI to make routing decisions independently
D) Reduce transparency regarding delivery prioritization to prevent challenges
A) Adjust the AI model to ensure equitable service distribution - AI should not unintentionally discriminate against certain areas.
A company uses AI to automate credit limit adjustments for customers. Low-income users report having their limits reduced despite no late payments. What governance response is appropriate?
A) Conduct an impact analysis and adjust AI decision criteria to prevent socioeconomic discrimination
B) Maintain the current model since it minimizes financial risk
C) Reduce AI transparency to protect trade secrets
D) Expand automation without addressing customer complaints
A) Conduct an impact analysis and adjust AI decision criteria to prevent socioeconomic discrimination - AI in finance should not disadvantage specific economic groups unfairly.
A major social media platform deploys an AI-driven content moderation system to identify and remove misinformation. However, a study finds that the system disproportionately flags posts written in certain non-English languages, leading to unfair content removal. Additionally, the moderation model is not interpretable, making it difficult for affected users to appeal decisions. What governance step should be taken?
A) Conduct a fairness audit, diversify training data, and introduce an appeal mechanism
B) Maintain AI-driven content moderation as it reduces misinformation
C) Reduce AI transparency to prevent backlash
D) Expand AI-driven content moderation without addressing linguistic fairness concerns
A) Conduct a fairness audit, diversify training data, and introduce an appeal mechanism - AI in content moderation must be fair and explainable.
What must AI providers do before placing a high-risk AI system on the EU market?
A) Perform a risk assessment, document AI system specifications, and ensure regulatory compliance
B) Deploy the AI system first and assess risks later
C) Require deployers to handle compliance independently
D) Keep AI risk assessment procedures confidential
A) Perform a risk assessment, document AI system specifications, and ensure regulatory compliance - AI providers are responsible for pre-market safety and compliance.
What is the role of AI accountability in governance frameworks?
A) Ensuring organizations take responsibility for AI-driven decisions and their consequences
B) Allowing AI to operate autonomously without legal or ethical constraints
C) Keeping AI decision-making confidential from regulatory bodies
D) Eliminating all human oversight in AI decision-making
A) Ensuring organizations take responsibility for AI-driven decisions and their consequences - AI accountability ensures ethical and legal responsibility for AI outcomes.
An AI-powered hiring tool is used by a corporation to screen candidates. After multiple complaints, an external review finds that rejected applicants are not given reasons for their rejection. According to AI governance best practices, what action should the company take?
A) Provide candidates with explainability reports detailing AI decision-making criteria
B) Maintain AI-driven hiring decisions since they optimize efficiency
C) Reduce AI transparency to prevent legal challenges
D) Expand AI hiring automation without addressing transparency concerns
A) Provide candidates with explainability reports detailing AI decision-making criteria - AI hiring tools should ensure transparency and fairness.
An e-commerce AI algorithm raises product prices for users who browse certain items multiple times. Some customers complain about unfair pricing. What governance action is most appropriate?
A) Implement price fairness guidelines to prevent exploitative pricing
B) Maintain dynamic pricing as it increases company revenue
C) Reduce transparency about pricing models to prevent customer complaints
D) Expand AI-driven pricing without addressing fairness concerns
A) Implement price fairness guidelines to prevent exploitative pricing - AI pricing strategies should not exploit user behavior unfairly.
Which governance practice is MOST effective for mitigating risks associated with AI model drift?
A) Deploying the model without post-deployment monitoring
B) Periodically retraining the model using updated, representative data
C) Ensuring the model remains unchanged after deployment
D) Ignoring model performance degradation
B) Periodically retraining the model using updated, representative data - Model drift occurs when data patterns change over time, requiring continuous monitoring and retraining.
Which regulation requires organizations to provide explanations for AI-driven decisions affecting individuals?
A) GDPR
B) CCPA
C) HIPAA
D) AI Transparency Act
A) GDPR - The General Data Protection Regulation mandates transparency and the right to explanation in AI decision-making.
A city government uses AI for predictive policing, but an independent study finds that low-income neighborhoods are disproportionately flagged as high-crime areas. What governance response should be prioritized?
A) Conduct an equity audit and recalibrate AI crime prediction models
B) Maintain AI-driven policing as it optimizes law enforcement
C) Reduce AI transparency to prevent community backlash
D) Expand AI crime prediction without addressing fairness concerns
A) Conduct an equity audit and recalibrate AI crime prediction models - AI in law enforcement should avoid reinforcing societal biases.
Which of the following is a challenge in balancing fairness and accuracy in AI models?
A) Adjusting AI decision-making without significantly reducing model performance
B) Ensuring AI models operate entirely without human oversight
C) Preventing regulatory bodies from scrutinizing AI fairness
D) Reducing the diversity of training datasets to improve efficiency
A) Adjusting AI decision-making without significantly reducing model performance - AI fairness interventions must balance equity with predictive accuracy.
An AI-driven news aggregation platform prioritizes emotionally charged content because it leads to higher engagement. Regulators warn that this may contribute to misinformation. What governance action should be taken?
A) Adjust the algorithm to prioritize credibility over engagement metrics
B) Continue using AI engagement-based ranking as it increases platform revenue
C) Remove human oversight to ensure AI neutrality
D) Reduce transparency in ranking factors to prevent external intervention
A) Adjust the algorithm to prioritize credibility over engagement metrics - AI-driven content distribution should promote factual information over sensationalism.
An AI-powered medical diagnosis system is found to be vulnerable to data poisoning, where manipulated training data leads to incorrect diagnoses. What governance measure should be implemented?
A) Strengthen data integrity measures and introduce model validation protocols
B) Maintain the AI diagnostic system since it improves efficiency
C) Reduce AI transparency to avoid exposing weaknesses
D) Automate all medical decisions without human review
A) Strengthen data integrity measures and introduce model validation protocols - Ensuring high-quality, tamper-free data prevents poisoning attacks.
An AI system that evaluates job applications is found to favor applicants with certain names over others. What should the company do?
A) Review training data and implement fairness constraints
B) Maintain the AI system as it follows historical hiring trends
C) Remove transparency in hiring decisions to prevent scrutiny
D) Allow AI to fully automate hiring without human intervention
A) Review training data and implement fairness constraints - AI hiring tools should not discriminate based on names or demographics.
A ride-sharing company implements AI-based surge pricing. A news investigation finds that fares during emergencies are disproportionately high. What governance measure should be prioritized?
A) Maintain surge pricing since it follows market demand
B) Implement ethical pricing caps during emergencies
C) Reduce transparency in pricing decisions to prevent regulatory scrutiny
D) Expand AI-driven pricing without public disclosure
B) Implement ethical pricing caps during emergencies - AI pricing models should prevent price exploitation in critical situations.
A predictive AI tool used for credit approvals is found to reject applications from minority groups at a disproportionately high rate. What is the best governance response?
A) Conduct an AI fairness audit and modify decision criteria
B) Maintain AI-driven credit risk assessments since they optimize financial risk
C) Reduce AI transparency to avoid regulatory intervention
D) Expand AI automation in lending without modifications
A) Conduct an AI fairness audit and modify decision criteria - AI in finance should not reinforce systemic bias.
An AI system designed for facial recognition in public spaces misidentifies individuals from a specific demographic at a significantly higher rate. Under AI governance best practices, what is the BEST course of action?
A) Ignore the discrepancy if overall accuracy is high
B) Increase the dataset size without considering demographic representation
C) Conduct a bias audit and retrain the model with a more diverse dataset
D) Restrict the use of the AI system to government agencies only
C) Conduct a bias audit and retrain the model with a more diverse dataset - Bias audits help ensure fairness and prevent systematic discrimination.
What is the role of human oversight in AI governance?
A) To intervene when AI decisions produce unfair or harmful outcomes
B) To ensure AI models operate without regulatory constraints
C) To minimize transparency in AI decision-making
D) To eliminate human influence in AI-driven processes
A) To intervene when AI decisions produce unfair or harmful outcomes - Human oversight ensures accountability and fairness in AI operations.
What is a key component of a successful AI governance strategy?
A) Aligning AI policies with ethical guidelines and regulatory requirements
B) Ensuring AI models remain proprietary and confidential
C) Eliminating transparency to avoid legal challenges
D) Allowing AI to function autonomously without oversight
A) Aligning AI policies with ethical guidelines and regulatory requirements - AI governance strategies must be ethical and legally compliant.
Why is explainability important in AI governance?
A) It allows users and regulators to understand how AI makes decisions
B) It ensures AI operates with minimal human oversight
C) It prioritizes proprietary secrecy over transparency
D) It maximizes AI computational efficiency without consideration for fairness
A) It allows users and regulators to understand how AI makes decisions - Explainability ensures AI decision-making can be interpreted and challenged.
What is a common challenge in mitigating AI bias?
A) AI models may learn unintended patterns from biased training data
B) AI automatically ensures fairness without intervention
C) AI models never need retraining once deployed
D) AI always produces unbiased outcomes
A) AI models may learn unintended patterns from biased training data - Fair AI governance requires ongoing bias detection and mitigation strategies.
An AI-driven job application screener in an EU company is flagged under regulatory review for discriminating against candidates based on nationality. What governance measure should be implemented?
A) Conduct an equity audit and adjust AI ranking algorithms
B) Maintain AI-driven hiring since it aligns with company policies
C) Reduce AI transparency to prevent legal challenges
D) Automate all job screening decisions without modifications
A) Conduct an equity audit and adjust AI ranking algorithms - The EU AI Act enforces non-discriminatory hiring practices for AI-driven systems.
A ride-sharing company’s AI-based surge pricing model results in disproportionately high fares for users in low-income areas. What governance response should the company take?
A) Implement fairness constraints to prevent discriminatory pricing
B) Maintain AI surge pricing since it follows market demand
C) Reduce transparency in AI pricing decisions to prevent public concerns
D) Expand AI-driven pricing without reviewing fairness issues
A) Implement fairness constraints to prevent discriminatory pricing - AI pricing strategies should be designed to avoid reinforcing social inequalities.
What is the purpose of a bias impact assessment in AI governance?
A) To identify and mitigate algorithmic bias before deployment
B) To increase the speed of AI decision-making
C) To replace human oversight in AI-driven decisions
D) To remove accountability from AI developers
A) To identify and mitigate algorithmic bias before deployment - Bias impact assessments help prevent unfair AI outcomes.
Which governance measure ensures that AI systems provide users with explanations for their decisions?
A) Explainability
B) Complexity
C) Efficiency
D) Algorithmic opacity
A) Explainability - AI transparency ensures that decisions are interpretable and can be questioned.
What is a key reason why post-processing fairness techniques are used in AI governance?
A) They adjust AI decision outcomes after model predictions to correct bias
B) They improve AI model efficiency by eliminating fairness constraints
C) They remove the need for transparency in AI decision-making
D) They allow AI systems to function without human intervention
C) They adjust AI decision outcomes after model predictions to correct bias - Post-processing techniques help improve fairness without modifying the underlying model.
A social media company deploys an AI-powered content moderation tool to detect harmful content. However, an audit finds that the system disproportionately flags posts from minority communities as policy violations. Further, affected users report that they are not given sufficient explanations for why their posts were removed. What governance action should be taken?
A) Conduct a fairness audit, retrain AI models, and implement a clear appeal mechanism
B) Maintain AI-driven moderation since it optimizes content filtering
C) Reduce AI transparency to prevent user complaints
D) Expand AI-driven content moderation without addressing fairness concerns
D) Conduct a fairness audit, retrain AI models, and implement a clear appeal mechanism - AI moderation should be unbiased and explainable.
Which of the following regulations is specifically designed to protect consumer privacy in California?
A) GDPR
B) HIPAA
C) CCPA
D) FTC Act
D) CCPA - The California Consumer Privacy Act (CCPA) establishes rules for consumer data privacy and AI governance.
What is the role of a human-in-the-loop system in AI governance?
A) Allowing AI systems to function autonomously without oversight
B) Enabling human intervention to review or override AI decisions
C) Increasing AI system processing speeds
D) Eliminating AI explainability concerns
B) Enabling human intervention to review or override AI decisions - Human-in-the-loop ensures critical decisions involve human oversight.
A multinational insurance company implements an AI-powered claims approval system. A regulatory audit reveals that the AI model consistently denies claims from lower-income regions at a higher rate, potentially due to biases in historical fraud detection data. The system also lacks explainability, making it difficult for affected customers to challenge denials. What governance step should the company prioritize?
A) Conduct an AI fairness audit and implement explainability mechanisms
B) Maintain current AI claims processing since it improves efficiency
C) Reduce transparency in AI decision-making to avoid reputational damage
D) Expand AI-driven claims automation without reviewing fairness concerns
A) Conduct an AI fairness audit and implement explainability mechanisms - AI in insurance must be transparent, unbiased, and accountable.
Which governance measure is essential for ensuring AI compliance with industry regulations?
A) Conducting periodic AI audits and compliance reviews
B) Reducing transparency in AI decision-making to prevent legal risks
C) Allowing AI systems to function without regulatory constraints
D) Keeping AI models confidential from regulators
A) Conducting periodic AI audits and compliance reviews - Regular audits help ensure AI compliance with ethical, legal, and industry standards.
A healthcare provider uses an AI model to predict patient risk scores for critical care allocation. A regulatory audit reveals that the AI model underestimates risk for minority patients due to an imbalance in the training data. What governance response should be prioritized?
A) Expand training datasets and ensure diverse representation in AI learning models
B) Maintain AI-driven patient risk assessments as they improve efficiency
C) Reduce AI transparency to prevent scrutiny
D) Automate all patient prioritization decisions without human review
A) Expand training datasets and ensure diverse representation in AI learning models - AI healthcare models must be trained on diverse, unbiased data.
Which of the following BEST describes the importance of AI governance in healthcare?
A) Ensuring patient safety and fairness in medical decision-making
B) Reducing the cost of AI development
C) Increasing AI automation to replace doctors
D) Preventing all medical errors
A) Ensuring patient safety and fairness in medical decision-making - AI governance ensures AI applications in healthcare are ethical, safe, and equitable.
Which of the following is a key principle of AI governance?
A) Ensuring AI systems align with ethical and regulatory standards
B) Maximizing AI performance without regulatory oversight
C) Reducing transparency to protect AI intellectual property
D) Allowing AI to operate without human intervention
A) Ensuring AI systems align with ethical and regulatory standards - AI governance ensures responsible and fair AI deployment.
An AI-driven insurance pricing model assigns higher premiums to certain customers based on their social media activity. Regulators question whether this approach is discriminatory. How should the insurance company respond?
A) Defend the model as a legitimate risk assessment tool
B) Conduct a fairness audit and review compliance with anti-discrimination laws
C) Increase reliance on AI decision-making to eliminate human bias
D) Reduce transparency to protect proprietary algorithms
B) Conduct a fairness audit and review compliance with anti-discrimination laws - AI in financial decision-making must align with ethical and legal fairness principles.
Which of the following best describes an unsupervised learning approach in AI?
A) AI models identify patterns and relationships in data without labeled outputs
B) AI models require explicit instructions for every decision
C) AI models only work with supervised training datasets
D) AI models cannot learn from raw data
B) AI models identify patterns and relationships in data without labeled outputs - Unsupervised learning finds hidden structures in data, such as in clustering.
A bank uses AI to automate mortgage lending. A study finds that the AI model rejects applications from minority groups at a higher rate. What governance measure should be taken?
A) Audit and retrain the AI system to prevent discriminatory lending
B) Maintain current lending decisions as they optimize financial risk
C) Reduce transparency in loan approval criteria to prevent scrutiny
D) Expand AI automation in lending without reviewing bias concerns
A) Audit and retrain the AI system to prevent discriminatory lending - AI financial tools should not reinforce systemic bias.
Which global AI governance principle ensures that AI systems respect human rights?
A) Fairness and non-discrimination
B) Data minimization
C) Algorithmic complexity
D) Proprietary secrecy
A) Fairness and non-discrimination - AI must be developed and used in a way that respects fundamental human rights and promotes equal opportunities.
Which organization role is responsible for overseeing AI governance implementation?
A) Chief AI Ethics Officer (CAIO) or AI Governance Lead
B) AI Development Engineers only
C) The IT department exclusively
D) AI governance requires no oversight
A) Chief AI Ethics Officer (CAIO) or AI Governance Lead - AI governance must be managed by dedicated roles within an organization.
What is the primary reason for requiring explainability in AI-driven healthcare diagnostics?
A) To increase AI model complexity for better accuracy
B) To ensure doctors and patients understand AI-based recommendations
C) To allow AI to function autonomously without human involvement
D) To prevent legal challenges by reducing AI accountability
B) To ensure doctors and patients understand AI-based recommendations - Explainability in AI improves trust and informed decision-making in healthcare.
A smart city deploys AI-driven energy distribution to reduce waste. However, lower-income areas experience more frequent power outages. What is the BEST response?
A) Adjust AI energy distribution to ensure equitable service
B) Maintain the current model since it optimizes energy efficiency
C) Reduce transparency in AI decision-making to prevent backlash
D) Expand AI automation in energy management without modification
A) Adjust AI energy distribution to ensure equitable service - AI governance must ensure fair allocation of public resources.
What is the purpose of bias audits in AI systems?
A) To identify and mitigate discriminatory patterns in AI decision-making
B) To make AI decision-making more complex and opaque
C) To ensure AI models operate autonomously without human oversight
D) To maximize AI computational efficiency
A) To identify and mitigate discriminatory patterns in AI decision-making - Bias audits help organizations detect and correct unfair AI behaviors.
What is the primary purpose of an AI impact assessment in governance frameworks?
A) To ensure AI systems operate without human intervention
B) To identify risks and mitigate potential societal harms of AI deployment
C) To maximize AI efficiency in decision-making
D) To reduce compliance costs for companies
B) To identify risks and mitigate potential societal harms of AI deployment - AI impact assessments help organizations manage risks proactively.
Which of the following is a key requirement under GDPR for AI systems that process personal data?
A) Users must have the right to explanation and contest AI decisions
B) AI systems must always operate autonomously
C) AI models must be proprietary and confidential
D) AI decisions cannot be legally challenged
A) Users must have the right to explanation and contest AI decisions - GDPR mandates transparency and accountability in AI decision-making.
A ride-sharing company’s AI dynamic pricing system increases fares disproportionately for users in lower-income neighborhoods. What governance measure should be implemented?
A) Maintain AI-driven pricing as it follows market dynamics
B) Introduce fairness constraints in AI pricing models
C) Reduce AI transparency in fare calculation to prevent customer complaints
D) Expand AI-driven fare adjustments without modification
C) Introduce fairness constraints in AI pricing models - AI pricing strategies should ensure equitable treatment of all users.
An AI-based loan approval system rejects a higher percentage of applications from minority groups compared to other applicants. What governance measure should be implemented?
A) Conduct a fairness audit and recalibrate the AI decision-making process
B) Maintain current AI decisions as they align with past lending trends
C) Reduce transparency in loan approvals to prevent regulatory scrutiny
D) Automate all loan decisions without human oversight
C) Conduct a fairness audit and recalibrate the AI decision-making process - AI financial models should be reviewed to prevent biased lending practices.
What is a key goal of explainability in AI models?
A) Enabling stakeholders to understand and challenge AI-driven decisions
B) Ensuring AI operates without human intervention
C) Maximizing algorithmic complexity to enhance decision-making
D) Eliminating the need for regulatory oversight
C) Enabling stakeholders to understand and challenge AI-driven decisions - Explainability enhances transparency and trust in AI.
A university deploys AI to screen scholarship applicants. A review finds that students from low-income schools are less likely to be shortlisted. What is the most ethical governance action?
A) Conduct an impact assessment and refine AI decision-making criteria
B) Maintain AI-driven scholarship screening as it improves efficiency
C) Reduce AI transparency to prevent public complaints
D) Automate all scholarship decisions without fairness audits
A) Conduct an impact assessment and refine AI decision-making criteria - AI in education should ensure fair access for all applicants.
Which of the following is a key focus of the NIST AI RMF?
A) Trustworthy and responsible AI development
B) Maximizing AI efficiency without considering risks
C) Eliminating human oversight in AI governance
D) Reducing transparency to protect proprietary AI models
A) Trustworthy and responsible AI development - The NIST RMF ensures AI systems remain ethical, transparent, and safe.
A credit-scoring AI system is found to have lower approval rates for individuals with unconventional employment histories. What is the most ethical governance response?
A) Audit and recalibrate the AI model to reduce unintended discrimination
B) Maintain the current system since it aligns with historical lending patterns
C) Reduce transparency in AI decision-making to prevent external challenges
D) Automate all credit assessments to remove human bias
A) Audit and recalibrate the AI model to reduce unintended discrimination - AI in finance should promote fair and unbiased decision-making.
A major social media platform deploys an AI-powered content moderation tool. A cybersecurity audit finds that the AI system is vulnerable to data extraction attacks, allowing attackers to reconstruct sensitive user data from model outputs. What governance action should be taken?
A) Implement stronger privacy-preserving AI techniques such as differential privacy
B) Maintain AI-driven content moderation as it improves user experience
C) Reduce AI transparency in decision-making to avoid regulatory scrutiny
D) Expand AI-driven moderation without additional security enhancements
C) Implement stronger privacy-preserving AI techniques such as differential privacy - AI content moderation must protect user privacy.
Which of the following is a critical risk of using AI for hiring decisions?
A) AI systems may reinforce bias in historical hiring patterns
B) AI eliminates discrimination in hiring decisions
C) AI hiring systems do not require audits
D) AI hiring is always more effective than human-based selection
A) AI systems may reinforce bias in historical hiring patterns - AI-driven recruitment should be monitored to prevent discrimination.
Why is it essential to include human oversight in high-risk AI decision-making systems?
A) To ensure that AI operates without any bias
B) To allow AI to function independently without regulatory interference
C) To provide accountability and intervention in case of errors or ethical concerns
D) To speed up AI decision-making by reducing human involvement
C) To provide accountability and intervention in case of errors or ethical concerns - Human oversight ensures responsible AI deployment in sensitive applications.
What is a key requirement for AI explainability in high-stakes decision-making?
A) Users must understand how AI reaches conclusions
B) AI decisions should be completely automated without human oversight
C) AI algorithms should remain confidential to protect company interests
D) AI systems must operate faster than traditional decision-making methods
A) Users must understand how AI reaches conclusions - Explainability ensures transparency and accountability in AI decision-making.
Which global AI principle ensures AI does not reinforce harmful biases?
A) Fairness and non-discrimination
B) Computational efficiency
C) Proprietary secrecy
D) Autonomous decision-making
A) Fairness and non-discrimination - AI should promote ethical and unbiased decision-making.
What is a key advantage of using explainable AI (XAI) in financial decision-making?
A) It allows regulators and customers to understand how AI makes lending decisions
B) It increases AI model complexity, making it harder to manipulate
C) It removes human oversight, ensuring faster decision-making
D) It ensures AI models remain completely free from bias
A) It allows regulators and customers to understand how AI makes lending decisions - Explainable AI improves transparency and trust in financial AI applications.
A hospital deploys an AI-based patient triage system that prioritizes patients based on survival likelihood. Critics argue that this disproportionately impacts older patients. What is the BEST governance response?
A) Review and recalibrate AI decision criteria to ensure ethical fairness
B) Maintain AI-driven triage since it optimizes hospital efficiency
C) Reduce transparency in AI decision-making to avoid controversy
D) Expand AI-based triage without regulatory review
A) Review and recalibrate AI decision criteria to ensure ethical fairness - AI in healthcare must ensure equitable treatment for all patients.
What is model drift in AI governance?
A) When an AI model’s performance degrades over time due to changing data patterns
B) When AI models become more transparent over time
C) When AI systems improve without human intervention
D) When AI models require less data for better accuracy
A) When an AI model’s performance degrades over time due to changing data patterns - Model drift can result in inaccurate predictions and must be managed through monitoring.
A self-driving AI system struggles with recognizing road signs in certain weather conditions. What governance step should be taken?
A) Improve training data and retrain AI models for adverse weather conditions
B) Maintain AI performance since it operates well in normal conditions
C) Reduce AI transparency about system weaknesses to avoid liability
D) Expand self-driving AI without additional testing
A) Improve training data and retrain AI models for adverse weather conditions - AI in transportation must function reliably across diverse conditions.
What is the role of the European Artificial Intelligence Board under the EU AI Act?
A) Coordinating AI regulatory enforcement and providing guidance across member states
B) Conducting conformity assessments for all AI systems
C) Deploying AI models in high-risk industries
D) Developing AI training data for providers
C) Coordinating AI regulatory enforcement and providing guidance across member states - The Board ensures harmonized application of AI governance rules.
What governance measure does the EU AI Act require for AI systems that impact fundamental rights?
A) Risk assessments and human oversight
B) Full automation without accountability
C) The ability to operate without audits
D) Complete exemption from legal challenges
A) Risk assessments and human oversight - AI systems impacting rights must be monitored for fairness and compliance.
What is the primary objective of AI governance?
A) To ensure AI aligns with ethical, legal, and regulatory standards
B) To maximize AI efficiency at all costs
C) To remove human oversight from AI decision-making
D) To eliminate all AI-related regulations
A) To ensure AI aligns with ethical, legal, and regulatory standards - AI governance establishes frameworks to promote responsible AI deployment.
Which of the following is a core requirement for AI explainability?
A) The ability for users to understand how AI decisions are made
B) The removal of all regulatory oversight from AI systems
C) The ability for AI to operate autonomously without explanation
D) The protection of AI decision-making as confidential information
A) The ability for users to understand how AI decisions are made - AI explainability ensures transparency and accountability.
Which of the following is a key characteristic of deep learning?
A) The use of multiple layers in neural networks to recognize complex patterns
B) The elimination of human intervention in all decision-making
C) The requirement that AI models use only structured data
D) The ability to function without labeled training data
C) The use of multiple layers in neural networks to recognize complex patterns - Deep learning models extract intricate features from raw data.
A retail company deploys an AI-based demand forecasting model to optimize inventory. However, smaller stores report frequent stock shortages due to the AI favoring high-traffic locations. Upon review, the model is found to prioritize profit-maximization over equitable stock distribution. What governance measure should be taken?
A) Adjust AI decision criteria to balance inventory allocation across all stores
B) Maintain AI-driven inventory predictions as they optimize profitability
C) Reduce AI transparency in decision-making to protect corporate interests
D) Expand AI automation in inventory management without modifications
C) Adjust AI decision criteria to balance inventory allocation across all stores - AI in supply chain management should ensure fairness in product distribution.
Which of the following scenarios presents the HIGHEST risk under AI governance frameworks?
A) An AI chatbot that provides general customer service responses
B) A self-learning AI system used for financial loan approvals
C) An AI-powered search engine ranking algorithm
D) An AI system that generates automated email responses
B) A self-learning AI system used for financial loan approvals - AI in financial services requires strict governance due to potential discrimination and regulatory impact.
An AI hiring tool consistently ranks candidates from elite universities higher than those from community colleges. What governance measure should be taken?
A) Maintain AI-driven rankings to preserve historical hiring patterns
B) Conduct a fairness audit and introduce diversity constraints
C) Reduce AI transparency in hiring decisions to prevent scrutiny
D) Automate all hiring decisions without modifications
B) Conduct a fairness audit and introduce diversity constraints - AI hiring systems should ensure fair evaluation of all candidates.
An AI-driven hiring platform is trained on past employee data and consistently selects candidates similar to existing staff, leading to a lack of diversity. What governance step should the company take?
A) Conduct a fairness audit and retrain the AI using a more diverse dataset
B) Maintain the current system as it is based on historical hiring success
C) Increase AI automation to remove human biases
D) Prevent external audits to protect proprietary hiring algorithms
A) Conduct a fairness audit and retrain the AI using a more diverse dataset - AI hiring tools must be monitored for bias and fairness to comply with regulations.
Which method is commonly used to mitigate bias in AI models?
A) Diverse and representative training data
B) Removing transparency in AI decision-making
C) Reducing accountability in AI governance frameworks
D) Ensuring AI operates without periodic fairness audits
A) Diverse and representative training data - AI fairness depends on training models with unbiased and inclusive data.
A hospital is using an AI model to prioritize emergency patients. However, it is found that the model unintentionally deprioritizes elderly patients. What governance measure should be taken?
A) Audit and retrain the model to ensure fair prioritization
B) Maintain current AI operations since they improve efficiency
C) Reduce AI transparency to avoid legal challenges
D) Automate all emergency room decisions without human intervention
A) Audit and retrain the model to ensure fair prioritization - AI in healthcare must be continuously evaluated for fairness and inclusivity.
A banking AI fraud detection model flags a disproportionately high number of transactions from migrant workers. What is the BEST governance response?
A) Conduct a bias audit and recalibrate the fraud detection algorithm
B) Maintain current fraud detection models as they reduce financial risk
C) Reduce AI transparency to prevent legal challenges
D) Expand AI fraud detection without reviewing for bias
A) Conduct a bias audit and recalibrate the fraud detection algorithm - AI financial systems must not disproportionately target specific demographic groups.
Which of the following is a key requirement under the EU AI Act for high-risk AI systems?
A) They must undergo regular compliance checks and risk assessments
B) They are exempt from regulatory oversight
C) They must be developed using open-source algorithms
D) They must be entirely autonomous
A) They must undergo regular compliance checks and risk assessments - High-risk AI systems require strict regulatory compliance.
An AI hiring platform consistently ranks male candidates higher for leadership positions compared to female applicants. What governance action should be taken?
A) Conduct a bias audit and recalibrate AI decision-making criteria
B) Maintain current AI hiring decisions as they align with industry norms
C) Reduce transparency in AI hiring to prevent regulatory scrutiny
D) Expand AI-driven hiring decisions without reviewing bias factors
A) Conduct a bias audit and recalibrate AI decision-making criteria - AI hiring tools should promote diversity and fair evaluations.
What is the main ethical challenge of AI-driven predictive policing?
A) The risk of reinforcing systemic biases in crime prediction
B) The elimination of crime in all communities
C) AI predictive policing requires no human intervention
D) AI-driven policing decisions are always more accurate than human decisions
A) The risk of reinforcing systemic biases in crime prediction - AI law enforcement tools should not disproportionately target specific populations.
A city implements an AI-based traffic management system to optimize rush-hour flow. After six months, traffic congestion worsens in low-income neighborhoods. What is the BEST governance response?
A) Maintain the system since overall citywide efficiency has improved
B) Conduct an equity assessment and adjust the model to ensure fair distribution of traffic benefits
C) Disable the AI system and return to manual traffic control
D) Expand the system to more neighborhoods without addressing current disparities
B) Conduct an equity assessment and adjust the model to ensure fair distribution of traffic benefits - AI in public infrastructure must be fair and equitable.
An AI system designed for facial recognition in public spaces misidentifies individuals from a specific demographic at a significantly higher rate. Under AI governance best practices, what is the BEST course of action?
A) Ignore the discrepancy if overall accuracy is high
B) Increase the dataset size without considering demographic representation
C) Conduct a bias audit and retrain the model with a more diverse dataset
D) Restrict the use of the AI system to government agencies only
C) Conduct a bias audit and retrain the model with a more diverse dataset - Bias audits help ensure fairness and prevent systematic discrimination.
Which of the following is the BEST way to ensure algorithmic fairness in AI models?
A) Regularly auditing training data and decision outputs for bias
B) Increasing the complexity of the AI model
C) Reducing human oversight to prevent subjective interference
D) Using only historical data without adjustments
A) Regularly auditing training data and decision outputs for bias - AI fairness requires continuous evaluation to prevent discriminatory outcomes.
A government agency uses AI to assess risk levels for public assistance programs. An independent study finds that lower-income applicants are disproportionately flagged as high risk. What is the most ethical governance response?
A) Conduct a fairness audit and recalibrate AI decision-making criteria
B) Maintain current AI risk assessment as it improves efficiency
C) Reduce AI transparency to prevent complaints from affected applicants
D) Expand AI-driven eligibility assessments without reviewing for fairness
A) Conduct a fairness audit and recalibrate AI decision-making criteria - AI governance should ensure fairness and equitable treatment for all applicants.
A bank uses an AI-driven credit approval system. A review finds that loan applicants from lower-income areas receive higher rejection rates. What is the best governance response?
A) Conduct a fairness audit and adjust AI decision-making criteria
B) Maintain AI credit approval models since they optimize financial risk
C) Reduce AI transparency in loan approval decisions
D) Expand AI-driven financial assessments without reviewing fairness concerns
C) Conduct a fairness audit and adjust AI decision-making criteria - AI in finance must ensure fair lending practices.
A company is developing an AI-powered recruitment tool, but an audit reveals that it favors candidates with traditional education over self-taught applicants. What is the BEST governance response?
A) Maintain the model since it aligns with past hiring trends
B) Conduct a fairness audit and adjust AI decision criteria
C) Reduce AI transparency to prevent regulatory challenges
D) Automate all hiring decisions to remove human subjectivity
B) Conduct a fairness audit and adjust AI decision criteria - AI hiring systems should ensure fair evaluation across all candidates.
A facial recognition AI used for security screening at airports has a higher false positive rate for travelers of certain ethnic backgrounds. What governance measure should be taken?
A) Conduct a bias audit and retrain the model with diverse demographic data
B) Maintain the current system since it improves security
C) Reduce AI transparency to prevent public criticism
D) Expand AI-based screening without addressing bias concerns
A) Conduct a bias audit and retrain the model with diverse demographic data - Facial recognition AI must be unbiased and fair in security applications.
A hospital uses AI to predict patient deterioration rates, but doctors report that the AI is overly aggressive in recommending ICU admissions. What is the BEST governance action?
A) Review and recalibrate the AI model to balance accuracy and clinical judgment
B) Maintain AI decision-making as it prioritizes patient safety
C) Reduce AI transparency to avoid challenges from doctors
D) Allow AI to fully automate ICU admission decisions
A) Review and recalibrate the AI model to balance accuracy and clinical judgment - AI in healthcare should complement, not replace, medical expertise.
What is the role of an AI ethics committee?
A) To oversee AI deployments and ensure compliance with ethical standards
B) To optimize AI for financial profitability
C) To reduce AI regulation for faster deployment
D) To ensure AI models remain proprietary
A) To oversee AI deployments and ensure compliance with ethical standards - Ethics committees help organizations develop responsible AI practices.
A self-driving vehicle company is designing its AI to prioritize safety. What ethical dilemma is MOST relevant to AI governance in autonomous vehicles?
A) Deciding how AI should react in unavoidable accident scenarios
B) Ensuring that AI reduces traffic congestion
C) Maximizing vehicle efficiency over human safety
D) Reducing transparency in decision-making to prevent liability
A) Deciding how AI should react in unavoidable accident scenarios - AI in transportation must be programmed with ethical decision-making frameworks.
A financial institution uses AI for loan approvals. A regulatory audit finds that the AI denies loans to applicants from specific zip codes at a higher rate. What governance measure should be taken?
A) Maintain the current model as it aligns with historical data
B) Conduct a fairness audit and adjust the AI model to prevent geographic bias
C) Reduce AI transparency to prevent regulatory intervention
D) Expand AI decision-making without modifying bias factors
B) Conduct a fairness audit and adjust the AI model to prevent geographic bias - AI in finance should not reinforce systemic discrimination.
Which of the following is a key factor in implementing a successful AI governance strategy?
A) Aligning AI development with organizational values and ethical guidelines
B) Allowing AI to operate without oversight for efficiency
C) Eliminating all regulations to speed up AI deployment
D) Keeping AI governance policies confidential to avoid scrutiny
A) Aligning AI development with organizational values and ethical guidelines - A well-defined AI governance strategy ensures responsible and ethical AI use.
A financial institution deploys an AI-powered chatbot to provide investment advice. However, users report that the chatbot gives inconsistent and sometimes risky recommendations. What is the BEST governance response?
A) Implement strict content validation rules and human oversight
B) Allow the chatbot to continue evolving its recommendations
C) Reduce transparency in AI decision-making to avoid regulatory issues
D) Expand the AI model without additional safeguards
A) Implement strict content validation rules and human oversight - AI in finance must ensure responsible and accurate recommendations.
Which principle is CRUCIAL in ensuring AI transparency?
A) Explainability
B) Black-box modeling
C) Model complexity
D) AI system latency
A) Explainability - AI transparency ensures users understand decision-making processes, increasing accountability and trust.
A government agency uses AI to identify high-risk areas for welfare fraud. Reports show that low-income neighborhoods are disproportionately targeted. What is the best governance response?
A) Conduct a fairness audit and recalibrate the AI fraud detection system
B) Maintain AI-driven fraud detection as it improves efficiency
C) Reduce AI transparency to prevent public concerns
D) Expand AI surveillance without adjusting for bias
C) Conduct a fairness audit and recalibrate the AI fraud detection system - AI in public services must be fair and unbiased.
A bank uses AI for automated loan approvals. A compliance audit under the EU AI Act finds that the model disproportionately denies loans to applicants from lower-income backgrounds. What governance step should be prioritized?
A) Conduct a fairness audit and adjust AI decision criteria
B) Maintain current AI-driven loan approvals as they optimize financial risk
C) Reduce transparency in AI-driven lending to prevent regulatory scrutiny
D) Expand AI automation in financial decisions without modifications
A) Conduct a fairness audit and adjust AI decision criteria - The EU AI Act requires fair and unbiased AI decision-making in financial services.
A company develops an AI-powered credit assessment tool that systematically underestimates the creditworthiness of applicants from certain socioeconomic backgrounds. Which action would BEST address this issue?
A) Increase model opacity to prevent external audits
B) Conduct fairness audits, improve dataset diversity, and apply bias correction techniques
C) Automate all credit assessments to remove human influence
D) Rely solely on historical financial data without adjustments
B) Conduct fairness audits, improve dataset diversity, and apply bias correction techniques - AI in finance must ensure fair and unbiased decision-making.
Which of the following best describes differential privacy in AI?
A) A technique that ensures AI models do not memorize individual data points
B) A method of making AI models more efficient
C) A strategy to eliminate bias in AI training data
D) A framework that removes human oversight from AI systems
A) A technique that ensures AI models do not memorize individual data points - Differential privacy is used to protect individual data while training AI models.
Which of the following is a requirement under the EU AI Act for high-risk AI applications?
A) They must undergo regular compliance checks and risk assessments
B) They are exempt from regulatory oversight
C) They must be entirely autonomous
D) They must be open-source
B) They must undergo regular compliance checks and risk assessments - High-risk AI systems require strict regulatory compliance.
Which regulation focuses on protecting personal data and privacy in AI systems within the European Union?
A) GDPR
B) CCPA
C) AI Fairness Act
D) HIPAA
A) GDPR - The General Data Protection Regulation mandates strict controls on AI data processing and privacy.
What is a primary reason for regulating AI under frameworks like the EU AI Act?
A) To promote unrestricted AI development
B) To prevent any form of AI automation
C) To ensure AI is safe, fair, and respects fundamental rights
D) To eliminate human oversight in AI decisions
C) To ensure AI is safe, fair, and respects fundamental rights - AI regulations mitigate risks associated with unfair or harmful AI outcomes.
Which governance framework provides global AI governance guidelines that emphasize human-centered AI?
A) OECD AI Principles
B) NIST AI Risk Management Framework
C) GDPR
D) CCPA
A) OECD AI Principles - These principles promote human-centered AI, transparency, and accountability.
What are the core functions of the NIST AI RMF?
A) Govern, Map, Measure, and Manage
B) Detect, Prevent, Resolve, and Report
C) Assess, Implement, Audit, and Optimize
D) Monitor, Evaluate, Distribute, and Analyze
A) Govern, Map, Measure, and Manage - These four functions guide organizations in assessing and mitigating AI-related risks.
A large e-commerce platform deploys an AI pricing algorithm that unintentionally increases prices for customers in low-income areas due to purchasing patterns. How should the company address this issue?
A) Maintain the pricing model since it follows market trends
B) Implement fairness constraints to prevent discriminatory pricing
C) Hide the pricing model to protect company trade secrets
D) Expand price increases to all customer segments to balance the impact
B) Implement fairness constraints to prevent discriminatory pricing - AI pricing should not result in unjust socioeconomic discrimination.
A city government deploys AI-driven facial recognition for public security. Civil rights groups raise concerns about potential mass surveillance and privacy violations. What governance step should be prioritized?
A) Implement strict privacy policies and allow independent audits
B) Maintain AI deployment since it improves security outcomes
C) Reduce transparency about AI data collection to prevent public scrutiny
D) Expand facial recognition to all public spaces without restrictions
A) Implement strict privacy policies and allow independent audits - AI surveillance must balance security needs with privacy rights.
Which of the following regulatory approaches is considered the MOST stringent for AI compliance?
A) Self-regulation by industry
B) Voluntary compliance frameworks
C) Government-mandated risk assessments and audits
D) No formal regulations
C) Government-mandated risk assessments and audits - This approach ensures compliance, accountability, and minimizes AI-related risks.
What is the primary purpose of the NIST AI Risk Management Framework (RMF)?
A) To provide guidelines for identifying, assessing, and mitigating AI risks
B) To remove regulatory oversight from AI decision-making
C) To accelerate AI deployment by bypassing risk assessments
D) To ensure AI systems remain confidential from public scrutiny
A) To provide guidelines for identifying, assessing, and mitigating AI risks - The NIST RMF provides a structured approach to AI risk management.
What is the primary function of an AI ethics committee within an organization?
A) To oversee AI deployment and ensure compliance with ethical and regulatory standards
B) To optimize AI performance by eliminating regulatory constraints
C) To remove human oversight from AI decision-making
D) To ensure AI models are kept confidential from external stakeholders
B) To oversee AI deployment and ensure compliance with ethical and regulatory standards - Ethics committees play a key role in AI governance and risk management.
Why is continuous AI monitoring necessary in governance frameworks?
A) To detect bias, security risks, and performance issues over time
B) To increase AI complexity without limitations
C) To remove the need for human oversight
D) To keep AI decision-making processes confidential
A) To detect bias, security risks, and performance issues over time - Ongoing AI monitoring ensures fairness and compliance with ethical standards.
A company is developing AI-powered surveillance drones for law enforcement. Privacy advocates express concerns about mass surveillance risks. What is the MOST ethical governance approach?
A) Implement strict data retention policies and public oversight mechanisms
B) Expand AI surveillance capabilities without public disclosure
C) Reduce transparency in AI decision-making to maintain security
D) Allow AI to fully automate surveillance decisions
A) Implement strict data retention policies and public oversight mechanisms - AI in surveillance must balance security with privacy rights.
What does the term ‘algorithmic accountability’ refer to in AI governance?
A) The responsibility of AI developers and organizations for AI outcomes
B) The process of making AI models faster and more efficient
C) The ability of AI to function independently of human input
D) The removal of regulatory oversight from AI decisions
A) The responsibility of AI developers and organizations for AI outcomes - Algorithmic accountability ensures responsible AI deployment.
A hospital deploys an AI system for diagnosing diseases. However, it has a significantly higher false negative rate for rare conditions. What governance step should be prioritized?
A) Maintain the AI system since it works well for common diseases
B) Improve training data diversity and retrain the AI model
C) Reduce AI transparency to prevent unnecessary patient concerns
D) Allow AI to fully automate diagnostic decisions without review
A) Improve training data diversity and retrain the AI model - AI healthcare systems must be accurate across all patient groups.
Which of the following best describes an adversarial attack in AI security?
A) A deliberate attempt to manipulate AI decisions by feeding misleading inputs
B) The ability of AI to detect and remove all biases from data
C) AI models that operate autonomously with minimal security concerns
D) An AI system that prioritizes efficiency over fairness
D) A deliberate attempt to manipulate AI decisions by feeding misleading inputs - Adversarial attacks exploit vulnerabilities in AI models.
A transportation company uses AI to optimize delivery routes, but the system prioritizes wealthier areas, leading to delays in lower-income neighborhoods. What is the BEST governance response?
A) Adjust the AI system to ensure equitable service distribution
B) Maintain current routing since it improves efficiency overall
C) Reduce transparency in AI decision-making to prevent complaints
D) Expand the system without addressing existing inequities
A) Adjust the AI system to ensure equitable service distribution - AI deployment in public services should be fair and unbiased.
A hiring AI system is found to disproportionately reject candidates with employment gaps, negatively impacting parents returning to work. What governance step should be taken?
A) Modify the AI model to ensure fairness for all applicants
B) Maintain current hiring algorithms as they reflect past hiring patterns
C) Remove human oversight to prevent subjectivity
D) Expand AI automation without adjusting bias factors
A) Modify the AI model to ensure fairness for all applicants - AI hiring should not disadvantage candidates based on career gaps.
What is the primary ethical concern with AI-based predictive policing models?
A) They increase police efficiency in crime prevention
B) They may reinforce systemic biases and unfairly target certain communities
C) They always provide accurate crime risk assessments
D) They require no human oversight in decision-making
B) They may reinforce systemic biases and unfairly target certain communities - AI in law enforcement must be carefully managed to prevent discrimination.
How does homomorphic encryption enhance AI security?
A) It allows computations to be performed on encrypted data without decryption
B) It removes the need for AI security monitoring
C) It reduces the transparency of AI decision-making
D) It ensures AI models operate without risk assessments
C) It allows computations to be performed on encrypted data without decryption - Homomorphic encryption preserves privacy while enabling secure AI processing.
Which AI governance principle is MOST critical when developing an AI system that predicts job candidate success?
A) Cost efficiency
B) Automation of all decision-making
C) Bias mitigation and fairness
D) Model complexity
C) Bias mitigation and fairness - AI hiring tools must ensure fair treatment and avoid discriminatory bias.
A ride-sharing company’s AI-based pricing model increases fares disproportionately for users in rural areas. What governance response should the company take?
A) Implement fairness constraints to prevent discriminatory pricing
B) Maintain AI surge pricing since it follows market demand
C) Reduce transparency in AI pricing decisions to prevent public concerns
D) Expand AI-driven pricing without reviewing fairness issues
A) Implement fairness constraints to prevent discriminatory pricing - AI pricing strategies should be designed to avoid reinforcing social disparities.
Which regulation governs AI-driven patient data privacy in the United States?
A) GDPR
B) HIPAA
C) CCPA
D) AI Fairness Act
B) HIPAA - The Health Insurance Portability and Accountability Act (HIPAA) regulates AI-driven healthcare data privacy in the U.S.
What is the primary ethical concern with AI in hiring processes?
A) AI models can reinforce biases from historical hiring data
B) AI hiring models are inherently unbiased
C) AI ensures faster hiring with no downsides
D) AI removes the need for fairness audits
A) AI models can reinforce biases from historical hiring data - AI must be carefully monitored to ensure fairness in recruitment decisions.
A financial institution deploys an AI model for automated loan approvals. After several months, analysts notice that applicants from certain ZIP codes are being denied loans at a significantly higher rate despite similar credit scores. What should the institution do to ensure compliance with AI fairness guidelines?
A) Ignore the issue if the model maintains high accuracy overall
B) Conduct an algorithmic audit to detect potential bias and retrain the model
C) Remove all geographic data to prevent bias without further investigation
D) Increase the model’s complexity to make decision-making less transparent
B) Conduct an algorithmic audit to detect potential bias and retrain the model - AI systems in financial services must be regularly audited for fairness and compliance.
Which AI governance framework is designed to promote responsible AI use within organizations?
A) AI governance policies and compliance structures
B) AI secrecy and proprietary restrictions
C) Eliminating AI transparency obligations
D) Preventing regulatory oversight of AI models
C) AI governance policies and compliance structures - Organizations need structured policies to ensure ethical and legal AI deployment.
What is the primary goal of fairness in AI governance?
A) To increase AI processing speeds
B) To ensure AI decisions do not disproportionately harm any group
C) To make AI systems completely independent of human input
D) To remove all decision-making biases
B) To ensure AI decisions do not disproportionately harm any group - Fairness ensures AI treats individuals and groups equitably.
A hospital AI system is designed to detect high-risk patients but consistently underestimates risks for certain ethnic groups. What governance measure should be implemented?
A) Improve training data diversity and retrain the AI model
B) Maintain the current model since it works well for most patients
C) Reduce AI transparency to avoid legal challenges
D) Automate all risk assessments to remove human bias
A) Improve training data diversity and retrain the AI model - AI in healthcare must provide accurate and fair assessments for all patients.
Which of the following best describes artificial intelligence (AI)?
A) The ability of machines to perform tasks that typically require human intelligence
B) A set of hardware components designed to enhance computing power
C) A regulatory framework designed to monitor human behavior
D) A system that operates purely based on fixed rules without learning
A) The ability of machines to perform tasks that typically require human intelligence - AI encompasses technologies like machine learning, natural language processing, and computer vision.
Which of the following best describes algorithmic bias?
A) AI system outputs that are consistently incorrect
B) Systematic and repeatable errors that unfairly favor or disadvantage certain groups
C) Slow AI performance due to insufficient training data
D) AI models that require human intervention for decision-making
B) Systematic and repeatable errors that unfairly favor or disadvantage certain groups - Algorithmic bias occurs when AI systems produce unfair or discriminatory outcomes.
What is the primary responsibility of the deployer of an AI system under the EU AI Act?
A) Ensuring the AI system is used according to regulatory requirements and monitoring its performance
B) Conducting pre-deployment risk assessments and documentation
C) Determining whether the AI system qualifies as high-risk
D) Granting market approval for AI systems
B) Ensuring the AI system is used according to regulatory requirements and monitoring its performance - Deployers are responsible for appropriate AI usage and post-market monitoring.
Under the EU AI Act, what is a mandatory requirement for high-risk AI systems?
A) They must undergo risk assessments and compliance checks
B) They must be entirely autonomous with no human oversight
C) They are exempt from legal accountability
D) They must be kept confidential from regulators
A) They must undergo risk assessments and compliance checks - High-risk AI applications require strict governance and oversight.
What is the purpose of algorithmic audits in AI governance?
A) To evaluate AI fairness, accuracy, and compliance
B) To increase AI efficiency at the cost of transparency
C) To remove the need for human oversight
D) To accelerate AI deployment without regulation
C) To evaluate AI fairness, accuracy, and compliance - Audits help organizations identify potential risks and biases in AI models.
Which OECD HUDERAF principle ensures that AI decision-making remains aligned with ethical and legal requirements?
A) Accountability
B) Universality
C) Diversity
D) Human-centeredness
A) Accountability - AI accountability ensures that organizations remain responsible for the ethical use of AI.
What role does encryption play in AI governance?
A) Protecting AI model data from unauthorized access
B) Ensuring AI operates without transparency
C) Reducing human oversight in AI decision-making
D) Eliminating explainability in AI processes
A) Protecting AI model data from unauthorized access - Encryption safeguards sensitive AI data and prevents security breaches.
What is a common risk associated with biased AI training data?
A) AI models may systematically disadvantage certain groups
B) AI will always make unbiased decisions
C) AI becomes more efficient without oversight
D) AI will never require retraining
A) AI models may systematically disadvantage certain groups - Biased training data can lead to unfair and discriminatory AI decisions.
A European financial institution plans to deploy an AI system for fraud detection. The system processes personal financial data and flags suspicious transactions. According to the EU AI Act, what is the provider’s responsibility before deploying the AI system?
A) Conduct a conformity assessment, document risk management procedures, and ensure compliance with transparency obligations
B) Deploy the AI model first and conduct compliance reviews only if a regulatory issue arises
C) Rely on deployers to manage all AI risk assessments and compliance obligations
D) Keep the AI fraud detection system’s methodology confidential
D) Conduct a conformity assessment, document risk management procedures, and ensure compliance with transparency obligations - AI providers must ensure high-risk AI systems meet legal requirements before deployment.
A ride-sharing company deploys an AI-based pricing model that increases fares during emergencies. Critics argue that this exploits vulnerable individuals. What governance step should be taken?
A) Introduce pricing caps during emergencies to prevent excessive charges
B) Maintain dynamic pricing to maximize company revenue
C) Reduce transparency in AI pricing decisions to prevent criticism
D) Expand surge pricing without public consultation
A) Introduce pricing caps during emergencies to prevent excessive charges - Ethical AI pricing models should avoid exploiting consumers in critical situations.
An AI model deployed by a government agency for benefit eligibility assessments is criticized for making incorrect denials. Applicants have no way to appeal AI decisions. Which governance measure should be prioritized?
A) Increase AI decision-making authority and remove human oversight
B) Implement an appeal process and ensure AI decisions are explainable
C) Reduce transparency to protect proprietary algorithms
D) Use social media sentiment analysis to validate AI outputs
B) Implement an appeal process and ensure AI decisions are explainable - AI governance must ensure fairness, accountability, and appealability in decision-making systems.
What is the purpose of AI explainability?
A) To allow users and regulators to understand AI decision-making processes
B) To improve AI efficiency at the cost of transparency
C) To reduce regulatory oversight on AI systems
D) To eliminate human involvement in AI operations
A) To allow users and regulators to understand AI decision-making processes - Explainability increases trust and accountability in AI governance.
A pharmaceutical company develops an AI model to predict drug side effects. However, the model struggles to identify risks for underrepresented populations. What governance measure should be prioritized?
A) Expand clinical trial datasets to include more diverse patient groups
B) Keep the AI system unchanged to maintain consistency
C) Automate all drug approval decisions using AI models
D) Reduce the regulatory oversight of AI-powered drug analysis
A) Expand clinical trial datasets to include more diverse patient groups - AI models in healthcare must be inclusive to ensure accurate predictions for all populations.
A financial institution integrates AI into its fraud detection and credit risk evaluation system. An independent review finds that the AI model disproportionately flags transactions from certain ethnic communities as suspicious. Furthermore, the AI uses a proprietary deep-learning model that lacks transparency, making it difficult for the institution to explain its risk evaluations to regulators. How should the institution address this issue?
A) Implement an AI explainability framework and conduct an audit to identify and mitigate bias
B) Maintain current AI fraud detection practices since they reduce financial risk
C) Reduce AI transparency in fraud detection criteria to avoid regulatory scrutiny
D) Expand AI-driven fraud detection without modifying risk assessment algorithms
C) Implement an AI explainability framework and conduct an audit to identify and mitigate bias - AI in finance must be explainable and unbiased.
A global e-commerce company uses an AI-powered recommendation engine to personalize product suggestions. A post-deployment analysis reveals that the system disproportionately pushes luxury items to higher-income users while systematically down-ranking budget-friendly products for lower-income users. Furthermore, the model lacks transparency, making it difficult to justify ranking decisions. What is the best governance action?
A) Conduct a fairness review and adjust AI ranking models to ensure equitable product exposure
B) Maintain AI-driven recommendations since they optimize revenue
C) Reduce AI transparency in ranking criteria to protect competitive advantage
D) Expand AI recommendation systems without reviewing fairness concerns
D) Conduct a fairness review and adjust AI ranking models to ensure equitable product exposure - AI in e-commerce should balance fairness and personalization.
What does HUDERAF stand for in the OECD AI governance framework?
A) Human-centeredness, Universality, Diversity, Explainability, Robustness, Accountability, Fairness
B) High Utility, Data Efficiency, Explainability, Responsibility, AI Fairness
C) Human Oversight, Data Ethics, Explainability, Responsibility, Automation Fairness
D) Human Decision-making, Utilization, Data Ethics, Responsibility, Algorithmic Fairness
A) Human-centeredness, Universality, Diversity, Explainability, Robustness, Accountability, Fairness - HUDERAF represents key OECD AI governance principles.
What is the main security risk of black-box AI models?
A) Lack of transparency, making it difficult to detect vulnerabilities
B) Excessive explainability, which reduces decision speed
C) AI requiring large datasets to function properly
D) AI being designed with fairness constraints
B) Lack of transparency, making it difficult to detect vulnerabilities - Black-box AI models hinder accountability and security auditing.
Which principle of OECD HUDERAF focuses on AI’s ability to withstand challenges and adversarial attacks?
A) Robustness
B) Diversity
C) Universality
D) Human-centeredness
A) Robustness - AI robustness ensures that systems remain reliable, secure, and resilient against threats.
Which of the following is a primary objective of AI risk management?
A) Identifying, assessing, and mitigating risks associated with AI deployment
B) Maximizing AI efficiency without considering potential risks
C) Eliminating human oversight to improve AI performance
D) Ensuring AI models operate without any regulatory scrutiny
A) Identifying, assessing, and mitigating risks associated with AI deployment - AI risk management helps prevent unintended consequences.
A government agency uses AI for public service eligibility assessments. Advocacy groups report that the system unfairly rejects applications from rural areas. What should the agency do?
A) Conduct an impact assessment and recalibrate the AI system
B) Maintain current eligibility rules to improve efficiency
C) Reduce AI transparency to prevent appeals
D) Expand AI decision-making without reviewing fairness concerns
A) Conduct an impact assessment and recalibrate the AI system - AI in public services must be equitable for all demographics.
An AI-powered loan approval system is found to reject applicants with lower credit history lengths at a significantly higher rate than other factors suggest is necessary for risk evaluation. What governance measure should be prioritized?
A) Review AI decision logic and adjust model weighting for credit history
B) Maintain AI-driven credit decisions since they reduce financial losses
C) Reduce transparency in AI lending decisions to prevent regulatory intervention
D) Automate all loan approvals without human oversight
D) Review AI decision logic and adjust model weighting for credit history - AI-based credit assessments should avoid unjustified biases.
A government agency deploys an AI-based biometric identification system for border control. A security analysis reveals that the system is vulnerable to spoofing attacks, where attackers use fake facial images to bypass authentication. What governance step should be taken?
A) Implement liveness detection and multi-factor authentication techniques
B) Maintain the current AI system since it improves efficiency
C) Reduce transparency in AI biometric decision-making
D) Expand AI-driven border control without additional security measures
D) Implement liveness detection and multi-factor authentication techniques - Security measures should prevent spoofing and unauthorized access.
A retail AI recommendation system consistently suggests high-priced items to users based on past purchasing behavior. Some users feel this is a form of price discrimination. What governance action is appropriate?
A) Implement diversity constraints in recommendations to provide more balanced options
B) Maintain the system as it maximizes revenue
C) Reduce AI transparency to prevent scrutiny
D) Expand the recommendation model without modifying its logic
A) Implement diversity constraints in recommendations to provide more balanced options - AI-driven consumer recommendations should not reinforce economic disparities.
An AI-driven fraud detection system at a bank flags transactions from small businesses at a disproportionately high rate. What governance action should be taken?
A) Maintain the current fraud detection model as it optimizes financial security
B) Conduct a bias audit and refine fraud detection criteria
C) Reduce transparency in fraud detection algorithms to prevent challenges
D) Expand AI-driven fraud detection without reviewing bias concerns
A) Conduct a bias audit and refine fraud detection criteria - AI fraud detection systems must be fair and non-discriminatory.
A government agency uses AI for welfare benefit assessments. An investigation reveals that the AI system disproportionately denies applications from single-parent households. What is the BEST corrective action?
A) Conduct a fairness audit and retrain the model with balanced data
B) Maintain the current system since it follows historical patterns
C) Remove human oversight to prevent subjective influence
D) Increase AI automation to improve efficiency
A) Conduct a fairness audit and retrain the model with balanced data - AI decision-making in public services must be fair and free from systemic bias.
What is the role of compliance monitoring in AI governance?
A) To ensure AI systems adhere to legal, ethical, and regulatory requirements
B) To accelerate AI deployment by removing regulatory constraints
C) To make AI models fully autonomous without oversight
D) To ensure AI decision-making remains confidential
A) To ensure AI systems adhere to legal, ethical, and regulatory requirements - Compliance monitoring ensures AI remains accountable and fair.
Which regulation primarily governs AI-driven healthcare decision-making and patient data privacy in the United States?
A) GDPR
B) HIPAA
C) CCPA
D) AI Fairness Act
B) HIPAA - The Health Insurance Portability and Accountability Act (HIPAA) regulates AI-driven healthcare decision-making and patient data privacy in the U.S.
A hospital AI system prioritizes younger patients for emergency treatment. Critics argue this disadvantages older patients. What is the BEST governance response?
A) Reassess the AI model and ensure fair treatment for all age groups
B) Maintain the current system since it optimizes hospital resources
C) Reduce transparency in patient prioritization to prevent legal issues
D) Automate all emergency treatment decisions without human oversight
A) Reassess the AI model and ensure fair treatment for all age groups - AI in healthcare should balance efficiency with ethical fairness.
A self-driving car manufacturer faces criticism after its AI prioritizes vehicle passengers over pedestrians in unavoidable accidents. What governance measure should the company implement?
A) Develop ethical decision-making guidelines and increase transparency
B) Maintain current AI behavior as it optimizes passenger safety
C) Reduce AI transparency to prevent legal liability
D) Expand AI automation without ethical oversight
A) Develop ethical decision-making guidelines and increase transparency - AI in autonomous vehicles must incorporate ethical decision frameworks.
A predictive AI tool is used to assess job candidates. A regulatory body finds that it consistently ranks applicants from certain socioeconomic backgrounds lower. What is the best governance response?
A) Conduct an algorithmic fairness audit and modify ranking criteria
B) Maintain current rankings since they align with past hiring trends
C) Reduce transparency in AI hiring decisions to avoid regulatory action
D) Expand AI decision-making in hiring without addressing bias concerns
A) Conduct an algorithmic fairness audit and modify ranking criteria - AI hiring tools should ensure fair evaluation across diverse candidates.
Which OECD HUDERAF principle ensures AI systems remain understandable and interpretable?
A) Explainability
B) Universality
C) Diversity
D) Robustness
A) Explainability - Explainability ensures that AI decisions can be interpreted and understood by users.
An AI hiring system is designed to select candidates based on past employee success patterns. However, it disproportionately rejects candidates from underrepresented groups. What is the BEST way to mitigate bias in this system?
A) Conduct a bias assessment, retrain the model with more diverse data, and introduce fairness constraints
B) Maintain the current selection criteria since it is based on historical data
C) Reduce transparency in AI decision-making to avoid legal challenges
D) Allow AI to fully automate hiring decisions without human intervention
A) Conduct a bias assessment, retrain the model with more diverse data, and introduce fairness constraints - AI hiring tools must prevent discrimination and ensure fairness.
A self-driving taxi company wants to optimize fuel efficiency. However, the AI model begins prioritizing routes that avoid lower-income neighborhoods. What is the BEST course of action?
A) Allow the AI to continue since it is maximizing efficiency
B) Modify the AI model to ensure equal service distribution across all communities
C) Reduce transparency about route selection to avoid regulatory scrutiny
D) Implement a pricing surcharge for affected areas to balance demand
B) Modify the AI model to ensure equal service distribution across all communities - AI-driven transportation must be fair and accessible to all.
What is the key purpose of AI risk assessments?
A) To identify, evaluate, and mitigate potential risks in AI deployment
B) To accelerate AI deployment by bypassing regulatory processes
C) To eliminate the need for AI governance frameworks
D) To make AI systems fully autonomous
A) To identify, evaluate, and mitigate potential risks in AI deployment - Risk assessments help prevent unintended consequences in AI applications.
What is a key risk of using AI for criminal sentencing recommendations?
A) AI ensures fair sentencing and removes bias
B) AI models may reinforce existing systemic biases in judicial systems
C) AI decision-making requires no human oversight in sentencing
D) AI-based sentencing decisions cannot be legally challenged
A) AI models may reinforce existing systemic biases in judicial systems - AI in criminal justice must be closely monitored for fairness.
What is the primary purpose of AI risk assessment in governance frameworks?
A) To identify and mitigate unintended consequences before deployment
B) To improve AI model speed and efficiency
C) To ensure AI systems operate without human intervention
D) To eliminate regulatory oversight
A) To identify and mitigate unintended consequences before deployment - AI risk assessments help organizations proactively manage potential ethical and legal risks.
Which of the following is a major challenge in ensuring fairness in AI-driven lending decisions?
A) AI models rely on historical financial data, which may contain systemic biases
B) AI lending decisions are always objective and fair
C) AI eliminates the need for human oversight in credit approval
D) AI-driven lending is immune to regulatory requirements
A) AI models rely on historical financial data, which may contain systemic biases - AI lending systems must be audited to ensure fair credit decisions.
Why is algorithmic impact assessment (AIA) important in AI governance?
A) It helps identify potential risks and societal impacts of AI deployment
B) It ensures AI models operate without human oversight
C) It maximizes AI efficiency at the cost of transparency
D) It allows companies to deploy AI without regulatory approval
A) It helps identify potential risks and societal impacts of AI deployment - AIAs ensure responsible AI usage and regulatory compliance.
Under the EU AI Act, which entity is primarily responsible for ensuring an AI system complies with conformity assessment procedures before deployment?
A) The provider of the AI system
B) The deployer of the AI system
C) The end user of the AI system
D) The regulatory oversight body
A) The provider of the AI system - The provider is responsible for conducting conformity assessments and ensuring compliance before deployment.
Under the EU AI Act, which entity is required to conduct conformity assessments for high-risk AI systems?
A) The AI provider
B) The deployer of the AI system
C) The national market surveillance authority
D) The European Data Protection Board
A) The AI provider - Providers must perform conformity assessments to ensure compliance before market deployment.
A university integrates AI into its admissions process to streamline applicant evaluation. However, after the first year of implementation, a review finds that applicants from rural schools are less likely to receive offers, despite comparable academic performance. Further investigation reveals that the AI model prioritizes historical admission trends, inadvertently reinforcing existing inequalities. How should the university address this issue?
A) Recalibrate AI decision criteria and conduct an equity audit to ensure fair admissions
B) Maintain AI-driven admissions since they improve efficiency
C) Reduce AI transparency to prevent challenges from rejected applicants
D) Expand AI-driven admissions without modifying ranking algorithms
C) Recalibrate AI decision criteria and conduct an equity audit to ensure fair admissions - AI in education must promote equitable access.
Which principle ensures organizations take responsibility for AI-driven outcomes?
A) Accountability
B) Complexity
C) Data secrecy
D) Automation
A) Accountability - AI accountability ensures that organizations are responsible for the impacts of AI decisions.
A company deploys an AI-powered resume screening tool. Female applicants receive lower rankings than male applicants with similar qualifications. What governance step should the company take?
A) Conduct an algorithmic audit and retrain the model with bias mitigation strategies
B) Maintain current hiring decisions to preserve company hiring trends
C) Reduce transparency in hiring decisions to avoid external scrutiny
D) Automate all hiring decisions without human intervention
A) Conduct an algorithmic audit and retrain the model with bias mitigation strategies - AI hiring tools should not reinforce gender-based discrimination.
A predictive policing AI system disproportionately assigns higher crime risk scores to individuals in certain neighborhoods. Local activists argue the system perpetuates racial bias. What is the BEST governance action?
A) Maintain the system’s current functionality since it improves crime prevention
B) Audit the training data and model outcomes for systemic bias and retrain the model
C) Expand AI surveillance in all neighborhoods to balance the dataset
D) Discontinue human oversight to prevent subjective interference
B) Audit the training data and model outcomes for systemic bias and retrain the model - AI-driven policing must be audited to prevent reinforcing historical biases.
Which AI governance principle ensures that AI systems remain accountable for their decisions?
A) Explainability
B) Autonomy
C) Complexity
D) Speed
C) Explainability - AI systems should provide understandable explanations for their decisions.
Which governance framework outlines principles for AI accountability and risk mitigation in the United States?
A) NIST AI Risk Management Framework (RMF)
B) GDPR
C) CCPA
D) OECD HUDERAF
B) NIST AI Risk Management Framework (RMF) - The NIST RMF helps organizations identify, assess, and mitigate AI risks.
A large healthcare provider deploys AI for disease diagnosis but finds that it misdiagnoses rare conditions more frequently than common illnesses. What governance measure should be prioritized?
A) Expand dataset diversity to include more rare disease cases
B) Reduce transparency in AI decision-making to prevent challenges
C) Maintain the current system since it is accurate for most cases
D) Automate all diagnosis decisions to remove human influence
A) Expand dataset diversity to include more rare disease cases - AI in healthcare must be trained on diverse and representative data.
Which of the following is a requirement under GDPR for AI systems processing personal data?
A) AI must provide a right to explanation for automated decisions
B) AI must operate entirely without human oversight
C) AI systems cannot be legally challenged
D) AI must prioritize efficiency over fairness
A) AI must provide a right to explanation for automated decisions - GDPR ensures transparency and accountability in AI decision-making.
A university deploys AI-driven admissions screening, but students from lower-income backgrounds receive lower acceptance rates. What governance action should be taken?
A) Conduct an algorithmic fairness audit and retrain the model
B) Maintain current admission decisions to ensure efficiency
C) Reduce AI transparency to prevent challenges from rejected applicants
D) Remove human oversight to ensure AI-driven objectivity
A) Conduct an algorithmic fairness audit and retrain the model - AI in education should ensure fair access to all applicants.
A multinational company uses AI-driven hiring software. An audit reveals that the model systematically prefers candidates from certain universities. What is the best governance response?
A) Conduct a bias audit and adjust AI decision criteria to ensure equal opportunity
B) Maintain current AI hiring practices since they align with business needs
C) Reduce AI transparency to avoid public scrutiny
D) Expand AI hiring automation without modifications
D) Conduct a bias audit and adjust AI decision criteria to ensure equal opportunity - AI hiring tools must prevent systemic discrimination.
An AI system used for employee performance evaluation consistently rates younger workers higher. What governance measure should be implemented?
A) Maintain AI performance evaluations since they optimize workforce efficiency
B) Conduct an age bias audit and adjust evaluation criteria
C) Reduce AI transparency to prevent employee complaints
D) Automate all performance assessments without modification
A) Conduct an age bias audit and adjust evaluation criteria - AI in workplace decision-making should be fair to all employees.
A company’s AI-based hiring tool consistently ranks younger candidates higher than older applicants. What governance measure should be taken?
A) Conduct a bias audit and recalibrate the AI model
B) Maintain AI hiring decisions since they align with company culture
C) Reduce transparency in hiring algorithms to prevent legal challenges
D) Automate all hiring decisions without human intervention
A) Conduct a bias audit and recalibrate the AI model - AI in hiring should ensure fair treatment for all age groups.
Under the EU AI Act, what is considered a high-risk AI system?
A) AI applications that significantly impact safety, rights, or livelihoods
B) AI models used solely for entertainment purposes
C) AI systems that do not require compliance monitoring
D) Any AI that operates without human oversight
A) AI applications that significantly impact safety, rights, or livelihoods - High-risk AI systems require strict governance and compliance under the EU AI Act.
A government agency uses AI to predict welfare fraud. Reports indicate that single-parent households are disproportionately flagged for investigation. What should be done?
A) Maintain AI fraud detection since it reduces financial risk
B) Reduce transparency in fraud detection models to avoid complaints
D) Expand AI-driven investigations without addressing fairness concerns
A) Conduct a fairness audit and recalibrate the fraud detection model
D) Conduct a fairness audit and recalibrate the fraud detection model - AI in public services must ensure fairness and avoid discrimination.
Which of the following is a major challenge of AI governance in hiring processes?
A) AI hiring models can reinforce biases present in historical data
B) AI completely eliminates discrimination in hiring decisions
C) AI hiring systems require no audits
D) AI hiring decisions are immune to legal challenges
B) AI hiring models can reinforce biases present in historical data - AI recruitment must be monitored to ensure fairness.
What is the core goal of the ‘Fairness’ principle in OECD HUDERAF?
A) Ensuring AI does not produce biased or discriminatory outcomes
B) Optimizing AI efficiency regardless of impact
C) Removing transparency obligations from AI models
D) Allowing AI models to operate without human oversight
A) Ensuring AI does not produce biased or discriminatory outcomes - Fairness in AI governance prevents social and economic biases.
Which AI security measure specifically helps mitigate adversarial attacks?
A) Adversarial training, which exposes AI models to attack-like inputs during training
B) Reducing AI transparency to protect proprietary algorithms
C) Eliminating human oversight in AI decision-making
D) Keeping AI models confidential to prevent public scrutiny
D) Adversarial training, which exposes AI models to attack-like inputs during training - This process strengthens AI robustness against adversarial threats.
An AI company develops a facial recognition system for law enforcement. Before placing the product on the EU market, the provider must:
A) Conduct a conformity assessment and ensure compliance with the EU AI Act
B) Deploy the system first and address compliance concerns afterward
C) Rely on law enforcement agencies to verify regulatory compliance
D) Keep the AI training data confidential to protect intellectual property
D) Conduct a conformity assessment and ensure compliance with the EU AI Act - High-risk AI applications must be assessed before market deployment.
A university deploys AI for student admissions. A report finds that first-generation college applicants are ranked lower. What is the most ethical governance action?
A) Conduct an equity audit and adjust AI decision criteria
B) Maintain AI admissions decisions as they improve efficiency
C) Reduce AI transparency to prevent complaints
D) Automate all admissions decisions without human oversight
A) Conduct an equity audit and adjust AI decision criteria - AI in education should ensure fair access to all applicants.
A law enforcement agency deploys AI facial recognition software, but an independent study finds it has a higher error rate for minority populations. What governance measure should be implemented?
A) Conduct a fairness audit and retrain AI with diverse datasets
B) Maintain AI-driven policing as it improves efficiency
C) Reduce AI transparency to avoid public criticism
D) Expand AI deployment without reviewing fairness concerns
B) Conduct a fairness audit and retrain AI with diverse datasets - AI in law enforcement must be fair and unbiased.
A large AI model is found to be using copyrighted material in its training data. How should the company using this AI system respond?
A) Conduct a data audit and remove unauthorized copyrighted material
B) Claim that AI-generated content is exempt from copyright laws
C) Increase model opacity to prevent detection
D) Limit the AI’s dataset to avoid generating creative outputs
A) Conduct a data audit and remove unauthorized copyrighted material - Ethical AI development requires compliance with copyright laws.
What is a key compliance requirement for high-risk AI systems under the EU AI Act?
A) Regular compliance audits and risk assessments
B) Full automation with no human oversight
C) Keeping AI decision-making confidential from regulators
D) Eliminating transparency to protect proprietary interests
A) Regular compliance audits and risk assessments - High-risk AI systems must be regularly evaluated for fairness, transparency, and accountability.
Which of the following best describes algorithmic bias?
A) Systematic errors in AI decision-making that disproportionately affect certain groups
B) AI systems that operate without any errors
C) AI algorithms that eliminate the need for human oversight
D) AI decision-making processes that always prioritize efficiency
A) Systematic errors in AI decision-making that disproportionately affect certain groups - Algorithmic bias can lead to discriminatory outcomes if not properly managed.
A government agency uses AI-driven facial recognition for law enforcement. An independent study finds that the system has a higher error rate for darker-skinned individuals. What governance step should be taken?
A) Conduct a fairness audit and retrain AI with diverse datasets
B) Maintain AI-driven facial recognition as it improves security
C) Reduce transparency in AI model performance to prevent public concern
D) Expand AI facial recognition without reviewing accuracy concerns
A) Conduct a fairness audit and retrain AI with diverse datasets - AI in law enforcement must be fair and accurate for all demographics.
A social media platform’s AI moderation tool is found to flag posts from non-native English speakers more frequently for misinformation. What is the most ethical governance response?
A) Conduct an equity audit and adjust AI content moderation algorithms
B) Maintain AI-driven moderation since it reduces misinformation overall
C) Reduce AI transparency in content moderation to prevent criticism
D) Expand AI-driven moderation without reviewing language bias
A) Conduct an equity audit and adjust AI content moderation algorithms - AI moderation must be fair and inclusive of diverse linguistic groups.
An AI-powered resume screening system consistently rejects candidates from certain universities. What governance measure should be implemented?
A) Review and adjust AI ranking criteria to ensure fairness
B) Maintain AI decisions as they reflect past hiring trends
C) Reduce transparency in AI hiring decisions to prevent scrutiny
D) Automate all hiring decisions without modifying selection criteria
A) Review and adjust AI ranking criteria to ensure fairness - AI hiring systems should not reinforce exclusionary hiring practices.
Which of the following is a key reason for implementing AI model documentation?
A) To provide transparency and accountability in AI development
B) To maximize AI efficiency by reducing oversight
C) To prevent AI models from being scrutinized by regulators
D) To ensure AI operates without human intervention
A) To provide transparency and accountability in AI development - Proper documentation helps regulators and stakeholders understand AI decision processes.
A major retailer implements AI-powered demand forecasting for inventory management, but the system consistently under-stocks products in lower-income neighborhoods. What governance measure should be implemented?
A) Conduct a fairness audit and refine AI decision-making processes
B) Maintain AI-driven inventory forecasts as they optimize logistics
C) Reduce AI transparency in supply chain management to avoid scrutiny
D) Expand AI-driven inventory management without reviewing fairness concerns
A) Conduct a fairness audit and refine AI decision-making processes - AI in supply chain management should ensure fair distribution.
An AI-powered recruitment tool consistently ranks candidates from elite universities higher than those from community colleges. What governance measure should be implemented?
A) Conduct a bias audit and adjust ranking criteria
B) Maintain AI rankings to reflect historical hiring patterns
C) Reduce AI transparency in hiring decisions to prevent scrutiny
D) Automate all hiring decisions without modifying AI models
A) Conduct a bias audit and adjust ranking criteria - AI hiring tools should ensure fair evaluations for all applicants.
A government agency uses AI to predict social benefits fraud. A review finds that the AI disproportionately flags single mothers. What is the most responsible governance action?
A) Maintain the AI model since it reduces fraud
B) Conduct a fairness audit and recalibrate the fraud detection algorithm
C) Reduce AI transparency to prevent complaints
D) Automate all fraud detection decisions without human oversight
B) Conduct a fairness audit and recalibrate the fraud detection algorithm - AI in public services must ensure fairness and avoid discrimination.
Which of the following is a requirement under the EU AI Act for high-risk AI applications?
A) Mandatory compliance with risk assessments and fairness evaluations
B) AI systems must be completely autonomous
C) AI models must not be used in high-risk domains
D) AI decisions cannot be legally challenged
A) Mandatory compliance with risk assessments and fairness evaluations - The EU AI Act imposes strict regulatory requirements on high-risk AI applications.
Which of the following is a key objective of AI risk management?
A) Identifying, assessing, and mitigating risks associated with AI deployment
B) Ensuring AI operates without any human intervention
C) Maximizing AI performance at the expense of compliance
D) Eliminating transparency in AI decision-making
A) Identifying, assessing, and mitigating risks associated with AI deployment - AI risk management helps prevent unintended consequences.
Which of the following best describes the concept of ‘human-in-the-loop’ AI?
A) AI systems that involve human oversight and decision-making in critical processes
B) AI that completely replaces human judgment in all areas
C) AI that operates without any human intervention
D) AI that is only used for entertainment purposes
A) AI systems that involve human oversight and decision-making in critical processes - Human-in-the-loop ensures human review in AI-driven decisions.
Which key principle of AI governance ensures responsibility for AI-related decisions?
A) Transparency
B) Accountability
C) Fairness
D) Automation
B) Accountability - Accountability ensures that AI systems are monitored, and responsibilities for decisions are clearly assigned.
Which of the following is a key principle of AI accountability?
A) Organizations must take responsibility for AI outcomes
B) AI should operate without human intervention
C) AI systems do not require audits
D) AI decision-making cannot be contested
A) Organizations must take responsibility for AI outcomes - AI accountability ensures responsible use and governance of AI technology.
What does AI accountability refer to in governance frameworks?
A) The responsibility of organizations for AI decisions and their consequences
B) The ability of AI to operate without legal constraints
C) The removal of human oversight from AI decision-making
D) The requirement that all AI models remain proprietary
A) The responsibility of organizations for AI decisions and their consequences - AI accountability ensures responsible use of AI technology.
A predictive healthcare AI model is designed to detect early signs of disease but frequently misdiagnoses minority patients. How should the healthcare provider respond?
A) Improve dataset diversity and retrain the model to enhance fairness
B) Maintain the current model since its overall accuracy is high
C) Reduce transparency in AI diagnostics to avoid public concern
D) Discontinue AI in healthcare due to potential ethical risks
A) Improve dataset diversity and retrain the model to enhance fairness - Healthcare AI must be inclusive and unbiased to ensure equal treatment for all patients.
A self-driving AI system is discovered to have a higher accident rate in urban environments with diverse pedestrian behaviors. What governance step should be prioritized?
A) Retrain the AI model with more diverse urban driving data
B) Maintain AI performance since it functions well in suburban areas
C) Reduce transparency in AI performance metrics to avoid liability
D) Allow AI to continue operating without additional testing
A) Retrain the AI model with more diverse urban driving data - AI in transportation must be tested across diverse environments for safety.
Which of the following is a best practice for enhancing AI transparency?
A) Providing clear documentation on AI decision-making processes
B) Reducing explainability to prevent regulatory challenges
C) Keeping AI model details confidential to protect intellectual property
D) Relying on fully autonomous AI decision-making
C) Providing clear documentation on AI decision-making processes - Transparent AI models help build trust and enable regulatory compliance.
An AI-powered job recruitment platform uses predictive analytics to rank candidates. After a regulatory investigation, it is found to favor candidates from wealthier backgrounds. What is the MOST ethical response?
A) Conduct a fairness audit and modify the algorithm to ensure equal opportunity
B) Maintain the model as it reflects historical hiring patterns
C) Remove human oversight to avoid subjective decision-making
D) Reduce transparency to prevent regulatory intervention
A) Conduct a fairness audit and modify the algorithm to ensure equal opportunity - AI hiring models must be designed to prevent socioeconomic discrimination.
A self-driving car company uses reinforcement learning to improve its vehicle’s decision-making. However, the system struggles in unpredictable real-world environments due to a lack of diverse training scenarios. What governance step should be taken?
A) Incorporate more diverse real-world driving scenarios into training data
B) Maintain AI-driven decision-making as it improves efficiency in most cases
C) Reduce AI transparency to protect proprietary information
D) Expand AI deployment without additional safety evaluations
D) Incorporate more diverse real-world driving scenarios into training data - Reinforcement learning should be exposed to a variety of conditions for robust decision-making.
Which fairness metric evaluates whether an AI model produces similar outcomes for different demographic groups?
A) Demographic parity
B) Model complexity
C) Data efficiency
D) Algorithmic transparency
D) Demographic parity - This metric assesses whether AI outputs are distributed fairly across different groups.
A self-driving car company deploys AI that prioritizes passenger safety over pedestrian safety in unavoidable accidents. What governance step should be taken?
A) Maintain AI decision-making as it optimizes passenger outcomes
B) Implement ethical decision-making frameworks and transparency guidelines
C) Reduce AI transparency to prevent public backlash
D) Expand AI automation in self-driving cars without modifications
C) Implement ethical decision-making frameworks and transparency guidelines - AI in transportation must incorporate ethical decision frameworks.
Which AI governance concept refers to organizations taking responsibility for AI-related outcomes?
A) Accountability
B) Efficiency
C) Automation
D) Data secrecy
A) Accountability - AI accountability ensures organizations remain responsible for AI decisions and their impacts.
What is the primary role of human oversight in AI decision-making?
A) To eliminate accountability for AI decisions
B) To intervene when AI outcomes produce unethical or unintended consequences
C) To make AI-driven processes completely autonomous
D) To reduce transparency in AI decision-making
B) To intervene when AI outcomes produce unethical or unintended consequences - Human oversight ensures responsible AI governance.
A government agency deploys an AI-driven welfare benefits system. Citizens report that they are being denied benefits without explanation, and the agency has no mechanism to provide reasons for rejections. What governance step should be taken?
A) Implement explainability mechanisms and provide affected individuals with reasoning behind AI decisions
B) Maintain AI-driven decisions since they improve efficiency
C) Reduce AI transparency to prevent administrative burdens
D) Expand AI automation without user appeals
D) Implement explainability mechanisms and provide affected individuals with reasoning behind AI decisions - AI in public services must be transparent and accountable.
A retail company wants to use an AI-powered recommendation engine to improve customer engagement. However, it notices that the algorithm tends to recommend products that reinforce existing socioeconomic divisions. How should the company address this issue?
A) Modify the recommendation algorithm to incorporate diversity constraints
B) Allow the algorithm to continue learning from past customer behavior
C) Remove explainability features to protect proprietary AI techniques
D) Limit access to the recommendation engine for lower-income users
A) Modify the recommendation algorithm to incorporate diversity constraints - AI recommendations should be optimized for fairness, not just engagement.
What is the purpose of the EU AI Act’s conformity assessments?
A) To ensure high-risk AI models comply with safety and ethical standards
B) To eliminate the need for human oversight in AI systems
C) To accelerate AI deployment by reducing regulatory constraints
D) To keep AI model compliance confidential
A) To ensure high-risk AI models comply with safety and ethical standards - Conformity assessments verify that AI systems adhere to EU regulations.
A multinational corporation deploys an AI-driven hiring platform. However, an audit reveals that the AI systematically favors applicants from certain elite universities, leading to a lack of diversity in hiring decisions. Further investigation shows that the bias originates from the historical hiring data used for training. What governance measure should be prioritized?
A) Conduct a bias audit and adjust AI decision-making criteria to promote diversity
B) Maintain AI-driven hiring decisions as they align with past recruitment trends
C) Reduce transparency in AI hiring decisions to prevent legal challenges
D) Expand AI hiring automation without modifications
D) Conduct a bias audit and adjust AI decision-making criteria to promote diversity - AI in recruitment should ensure fair evaluation for all applicants.
An AI company is deploying an algorithm for predictive policing. To align with ethical AI principles, what should be the MOST important priority?
A) Ensuring the system maximizes efficiency in identifying high-risk areas
B) Training the model on existing crime data without reviewing for bias
C) Conducting fairness audits to prevent racial or socioeconomic discrimination
D) Using a proprietary black-box model to protect competitive advantage
C) Conducting fairness audits to prevent racial or socioeconomic discrimination - Predictive policing AI must be audited to avoid biased enforcement and discriminatory impact.
What is a key requirement for AI fairness in hiring algorithms?
A) AI must not favor any specific demographic group
B) AI must only be used for executive-level hiring
C) AI must operate without human oversight
D) AI must base all hiring decisions purely on historical data
A) AI must not favor any specific demographic group - AI hiring systems should promote fair and equal opportunities.
Which legal framework is most relevant to AI data privacy and user consent in Europe?
A) CCPA
B) GDPR
C) HIPAA
D) FTC Act
B) GDPR - The General Data Protection Regulation (GDPR) sets strict guidelines on AI data privacy, user consent, and accountability.
A retail company implements AI-powered inventory management, but store managers report that the system consistently under-stocks high-demand products in urban areas. What governance measure should be implemented?
A) Conduct an AI model audit and refine inventory decision-making processes
B) Maintain AI inventory predictions as they optimize company logistics
C) Reduce transparency in AI-driven supply chain management to avoid scrutiny
D) Expand AI inventory management without addressing concerns
A) Conduct an AI model audit and refine inventory decision-making processes - AI in supply chain management should be continuously monitored for accuracy and fairness.
Which governance principle is MOST important for AI used in criminal sentencing recommendations?
A) Transparency and explainability
B) AI system efficiency
C) Automation without human intervention
D) Data minimization
A) Transparency and explainability - AI in criminal justice must be interpretable to ensure fair and accountable decision-making.
A healthcare provider deploys an AI diagnostic tool classified as high-risk. During use, medical staff report a high false-negative rate for certain conditions. Under the EU AI Act, what is the deployer’s responsibility?
A) Notify the provider and the relevant market surveillance authority while reviewing AI system use
B) Continue using the AI model while waiting for regulatory updates
C) Disable explainability features to avoid liability concerns
D) Reduce transparency in AI decisions to protect confidential medical records
B) Notify the provider and the relevant market surveillance authority while reviewing AI system use - Deployers must report safety concerns and address system risks.
A hospital deploys an AI-based diagnostic tool but finds that it has a significantly lower accuracy rate for diagnosing rare diseases in minority populations. What governance action should be prioritized?
A) Expand training datasets to ensure diversity in AI learning models
B) Maintain AI diagnostic decisions since they improve efficiency
C) Reduce transparency in AI medical decision-making to prevent public concern
D) Automate all AI-driven diagnostics without human review
A) Expand training datasets to ensure diversity in AI learning models - AI healthcare tools must be inclusive and accurate for all demographics.
What obligation do AI providers have when placing a high-risk AI system on the EU market?
A) They must establish a risk management system, maintain technical documentation, and conduct conformity assessments
B) They are only required to notify regulators post-deployment
C) They must delegate all responsibility to AI deployers
D) They need to ensure AI is proprietary and remains confidential
B) They must establish a risk management system, maintain technical documentation, and conduct conformity assessments - These obligations ensure compliance before market release.
A predictive policing AI disproportionately flags low-income neighborhoods as high-risk areas. What governance measure should be taken?
A) Conduct a fairness audit and recalibrate AI decision-making criteria
B) Maintain AI policing decisions as they optimize law enforcement
C) Reduce transparency in AI crime prediction to prevent backlash
D) Expand AI-driven policing without reviewing bias concerns
A) Conduct a fairness audit and recalibrate AI decision-making criteria - AI in law enforcement must ensure fairness and avoid reinforcing systemic biases.
What is the purpose of algorithmic impact assessments in AI governance?
A) To identify potential ethical and legal risks before AI deployment
B) To increase AI efficiency without regulatory concerns
C) To automate all AI decision-making
D) To eliminate AI bias completely
A) To identify potential ethical and legal risks before AI deployment - Algorithmic impact assessments help organizations address risks proactively.
Which AI governance principle focuses on reducing harm and ensuring system resilience?
A) Privacy by design
B) Accountability
C) Robustness and security
D) AI automation
C) Robustness and security - Ensuring AI systems are secure and resilient to attacks helps minimize unintended harm.
A multinational corporation wants to deploy an AI-powered hiring tool globally. However, different countries have varying laws on AI-driven hiring decisions. Which AI governance challenge is MOST relevant in this case?
A) Cross-border regulatory compliance
B) AI system scalability
C) Reducing algorithmic complexity
D) Increasing AI autonomy
A) Cross-border regulatory compliance - AI systems must comply with diverse international legal frameworks when deployed globally.
A university implements an AI system for student performance predictions. Faculty members argue that the AI disproportionately predicts lower success rates for students from certain socioeconomic backgrounds. What is the most ethical response?
A) Review and adjust the AI’s training data and algorithmic fairness constraints
B) Remove human faculty oversight to ensure AI independence
C) Maintain AI-generated predictions as final decisions without appeal
D) Reduce transparency to protect AI model confidentiality
A) Review and adjust the AI’s training data and algorithmic fairness constraints - AI in education should support fairness and equal opportunities for all students.
Which of the following is a best practice for implementing AI governance in an enterprise setting?
A) Establishing clear AI policies, accountability structures, and compliance mechanisms
B) Keeping AI systems confidential to protect proprietary interests
C) Reducing transparency to avoid regulatory intervention
D) Removing all human oversight to improve efficiency
C) Establishing clear AI policies, accountability structures, and compliance mechanisms - AI governance requires structured policies and oversight.
What is the main goal of AI regulatory compliance frameworks?
A) To ensure AI systems adhere to ethical, legal, and security standards
B) To eliminate all restrictions on AI development
C) To speed up AI deployment without safety measures
D) To remove human oversight from AI decision-making
A) To ensure AI systems adhere to ethical, legal, and security standards - Compliance frameworks help mitigate AI-related risks.
Which principle of the NIST AI RMF ensures AI systems remain adaptable to new threats and risks?
A) Continuous monitoring and assessment
B) Eliminating transparency in AI models
C) Allowing AI to operate without audits
D) Reducing human oversight
A) Continuous monitoring and assessment - AI systems require ongoing evaluation to remain secure and trustworthy.
A self-driving car manufacturer deploys an AI model for collision avoidance. However, reports emerge that the system fails to detect pedestrians who use wheelchairs due to biases in the training data. What governance measure should be taken?
A) Improve AI training data diversity and retrain the model with inclusive datasets
B) Maintain AI-driven safety decisions since they optimize overall traffic safety
C) Reduce AI transparency in pedestrian detection to prevent liability concerns
D) Expand self-driving AI deployment without addressing bias concerns
C) Improve AI training data diversity and retrain the model with inclusive datasets - AI in transportation should prioritize inclusivity and safety.
A healthcare AI system prioritizes younger patients for organ transplants due to longer life expectancy. What governance principle is at stake?
A) Fairness and non-discrimination
B) AI model efficiency
C) Data minimization
D) AI automation
A) Fairness and non-discrimination - AI healthcare decisions must be ethically justified and avoid unjust discrimination.
What is the primary goal of the EU AI Act?
A) To establish a regulatory framework for AI that prioritizes safety and fundamental rights
B) To ban all high-risk AI applications
C) To eliminate AI bias completely
D) To allow AI models to operate without oversight
A) To establish a regulatory framework for AI that prioritizes safety and fundamental rights - The EU AI Act ensures AI development and deployment align with ethical standards.
An AI system used in online content moderation starts disproportionately flagging posts from specific political groups as misinformation. Which governance measure would BEST help address this issue?
A) Implement human oversight and transparency reporting on moderation decisions
B) Allow the AI model to continue learning from user interactions without intervention
C) Increase the model’s reliance on keyword detection rather than context analysis
D) Reduce transparency in moderation policies to prevent public scrutiny
A) Implement human oversight and transparency reporting on moderation decisions - AI moderation must balance accuracy with fairness and transparency.
A predictive AI system used for credit risk assessment assigns lower scores to applicants from specific zip codes. What governance measure should be taken?
A) Conduct a bias audit and remove geographical bias from decision-making
B) Maintain AI-driven credit scoring as it optimizes financial risk assessment
C) Reduce transparency in AI credit scoring to avoid regulatory scrutiny
D) Automate all credit decisions without reviewing fairness concerns
A) Conduct a bias audit and remove geographical bias from decision-making - AI in finance must not contribute to redlining or economic discrimination.
A financial AI system that assesses credit risk denies a disproportionately high percentage of loan applications from minority communities. What is the best governance action?
A) Conduct a bias audit and adjust AI decision criteria
B) Maintain AI-driven credit risk assessments since they optimize financial risk
C) Reduce AI transparency to prevent regulatory intervention
D) Expand AI automation in lending without modifications
A) Conduct a bias audit and adjust AI decision criteria - AI financial models should be audited regularly to prevent discriminatory lending practices.
An AI-driven credit scoring system unintentionally denies loans to individuals in certain geographic areas due to historical data patterns. This is an example of:
A) Algorithmic transparency
B) Disparate impact
C) Model robustness
D) AI optimization
B) Disparate impact - AI decisions can unintentionally disadvantage certain groups even without explicit bias in programming.
A city government in the EU deploys AI-driven surveillance for public safety. Civil rights groups raise concerns about mass surveillance risks. What governance response is required under the EU AI Act?
A) Conduct a privacy impact assessment and ensure regulatory compliance
B) Maintain AI surveillance as it improves law enforcement efficiency
C) Reduce transparency in AI-driven surveillance to avoid public opposition
D) Expand AI-driven surveillance without reviewing privacy risks
A) Conduct a privacy impact assessment and ensure regulatory compliance - The EU AI Act mandates strict oversight for biometric and surveillance AI.
A predictive policing AI flags individuals as high-risk offenders based on neighborhood crime statistics. Critics argue this reinforces systemic bias. What is the BEST governance response?
A) Conduct an audit and retrain the model with fairness constraints
B) Maintain the AI system as it improves law enforcement efficiency
C) Reduce public access to crime prediction data to prevent scrutiny
D) Expand AI enforcement without modifying risk assessments
A) Conduct an audit and retrain the model with fairness constraints - AI policing should not reinforce societal biases or target specific communities unfairly.
A banking AI system is trained on historical financial data but is found to reject loan applications from younger applicants at a much higher rate. What is the BEST governance response?
A) Conduct a bias assessment and introduce fairness constraints in the model
B) Allow the model to continue operating since it reflects historical financial trends
C) Reduce transparency in AI decision-making to protect the bank’s reputation
D) Increase reliance on historical data to ensure consistency
A) Conduct a bias assessment and introduce fairness constraints in the model - AI financial systems should be fair and free from age discrimination.
A multinational corporation deploys an AI-powered recruitment platform. A compliance review finds that the AI consistently favors candidates from certain elite universities while under-ranking applicants from diverse educational backgrounds. How should the organization respond?
A) Conduct a bias audit and recalibrate AI decision-making criteria
B) Maintain AI hiring decisions as they align with past recruitment patterns
C) Reduce transparency in AI-driven hiring decisions to prevent legal challenges
D) Expand AI hiring automation without modifying ranking algorithms
C) Conduct a bias audit and recalibrate AI decision-making criteria - AI in hiring should ensure equal opportunity for all candidates.
An AI company develops a high-risk medical diagnosis system. The notified body conducting the conformity assessment finds that the model lacks sufficient documentation regarding bias mitigation strategies. What must the provider do before receiving market approval?
A) Implement bias mitigation techniques, update documentation, and resubmit for assessment
B) Deploy the AI system and update documentation at a later stage
C) Rely on deployers to determine if bias mitigation is necessary
D) Keep AI training methodologies confidential to protect proprietary interests
A) Implement bias mitigation techniques, update documentation, and resubmit for assessment - Conformity assessments require full documentation before AI deployment.
A retail company uses AI for customer profiling, but the system disproportionately offers discounts to wealthier customers. What governance step should be taken?
A) Adjust AI decision-making to ensure fair treatment across all income levels
B) Maintain current AI pricing strategies as they optimize company profits
C) Reduce transparency in customer segmentation to prevent backlash
D) Expand AI-driven customer profiling without modifying fairness concerns
A) Adjust AI decision-making to ensure fair treatment across all income levels - AI pricing models should be equitable and not exploit customer segmentation biases.
What is the purpose of algorithmic auditing in AI governance?
A) To identify and mitigate risks such as bias and unfair outcomes
B) To increase AI efficiency without considering ethical concerns
C) To replace human oversight with fully autonomous AI decision-making
D) To limit AI adoption across industries
A) To identify and mitigate risks such as bias and unfair outcomes - AI audits help ensure ethical and fair AI deployment.
What is the role of continuous monitoring in AI governance?
A) Ensuring AI models are regularly audited and adjusted to remain compliant
B) Eliminating the need for human oversight in AI decision-making
C) Ensuring AI operates autonomously without regulatory interference
D) Keeping AI governance frameworks static to prevent disruptions
D) Ensuring AI models are regularly audited and adjusted to remain compliant - AI systems require continuous assessment to mitigate risks.
What is the importance of AI transparency?
A) It allows stakeholders to understand and challenge AI decisions
B) It ensures AI operates without any human intervention
C) It allows AI systems to function autonomously without documentation
D) It removes the need for regulatory compliance
A) It allows stakeholders to understand and challenge AI decisions - Transparency is essential for AI governance and accountability.
Which type of bias occurs when an AI system learns patterns from historical data that reinforce social inequalities?
A) Historical bias
B) Selection bias
C) Automation bias
D) Proxy bias
A) Historical bias - This occurs when AI models replicate existing biases present in historical data.
Which governance measure ensures AI decision-making is understandable and explainable to stakeholders?
A) Transparency
B) Efficiency
C) Algorithmic secrecy
D) Proprietary protection
A) Transparency - Transparent AI systems allow stakeholders to understand and challenge AI-driven decisions.
Who is responsible for maintaining technical documentation for high-risk AI systems under the EU AI Act?
A) The AI provider
B) The deployer
C) The European Artificial Intelligence Board
D) The notified body
B) The AI provider - Providers must maintain detailed records of AI system specifications and risk management.
A university uses an AI model for student admissions, which favors applicants from wealthier backgrounds. What governance step should be prioritized?
A) Conduct a fairness audit and implement bias mitigation techniques
B) Maintain current AI admissions as they align with past data trends
C) Reduce transparency in admissions criteria to prevent external scrutiny
D) Expand AI-driven admissions without adjusting bias factors
A) Conduct a fairness audit and implement bias mitigation techniques - AI in education should promote equal access and opportunity.
Which AI principle ensures that AI systems remain robust against errors and unexpected conditions?
A) Resilience
B) Complexity
C) Transparency
D) Accountability
A) Resilience - AI resilience ensures that models function reliably in diverse environments and resist failures.
Under the EU AI Act, which specific requirement applies to biometric identification systems in public spaces?
A) They must undergo prior authorization and meet strict necessity and proportionality tests
B) They are exempt from regulatory scrutiny due to national security concerns
C) They must operate without explainability obligations to ensure efficiency
D) They are classified as low-risk AI applications and require minimal oversight
A) They must undergo prior authorization and meet strict necessity and proportionality tests - The EU AI Act places stringent controls on biometric AI systems used in public spaces.
A city government deploys AI traffic management that optimizes congestion, but it increases traffic in lower-income areas. What governance step should be prioritized?
A) Adjust AI algorithms to ensure equitable traffic distribution
B) Maintain current traffic patterns since overall efficiency improves
C) Reduce transparency in AI-driven decisions to prevent criticism
D) Expand AI-driven traffic management without modification
A) Adjust AI algorithms to ensure equitable traffic distribution - AI public services must not disproportionately impact specific communities.
Which of the following best describes a high-risk AI system under the EU AI Act?
A) AI applications that significantly affect individuals’ rights, health, or safety
B) AI systems that operate without bias or risk
C) AI used purely for entertainment and media applications
D) AI models that are exempt from compliance checks
A) AI applications that significantly affect individuals’ rights, health, or safety - High-risk AI systems require strict governance and monitoring.
Which of the following is a security concern related to AI governance?
A) Adversarial attacks that manipulate AI model outputs
B) AI’s ability to process large datasets
C) AI improving efficiency in business operations
D) AI transparency and explainability
D) Adversarial attacks that manipulate AI model outputs - AI security measures must prevent malicious manipulations that distort decision-making.
A hospital deploys an AI model to prioritize patients for organ transplants. It is found that wealthier patients receive higher priority regardless of medical urgency. What governance measure should be implemented?
A) Recalibrate AI decision criteria to ensure fair treatment for all patients
B) Maintain AI prioritization since it aligns with hospital policies
C) Reduce AI transparency in medical decisions to avoid scrutiny
D) Automate all transplant decisions without human oversight
A) Recalibrate AI decision criteria to ensure fair treatment for all patients - AI in healthcare should be equitable and prioritize based on medical urgency.
A deployer integrates an AI-powered recruitment system into their hiring process. A post-deployment audit reveals that the model discriminates against applicants from underrepresented backgrounds. What must the deployer do according to the EU AI Act?
A) Report the issue to regulatory authorities and collaborate with the AI provider for corrective actions
B) Maintain the AI-driven hiring process as it improves efficiency
C) Reduce AI transparency to prevent reputational risk
D) Expand AI hiring automation without addressing bias concerns
A) Report the issue to regulatory authorities and collaborate with the AI provider for corrective actions - Deployers must ensure AI systems comply with fairness regulations.
Which principle of AI governance ensures that AI users and stakeholders understand how decisions are made?
A) Transparency
B) Efficiency
C) Data secrecy
D) Algorithmic opacity
C) Transparency - AI decision-making should be interpretable and accessible to stakeholders.
An AI software company is preparing to introduce an automated resume screening system in the EU. The company is unsure whether their product qualifies as a high-risk AI system. What governance step should the company take first?
A) Perform a risk classification assessment based on EU AI Act guidelines
B) Deploy the AI model and wait for regulatory feedback
C) Restrict AI explainability to protect proprietary algorithms
D) Rely on end users to determine if compliance measures are necessary
B) Perform a risk classification assessment based on EU AI Act guidelines - AI providers must determine whether their system falls under high-risk regulations.
An AI company is deploying an algorithm for predictive policing. To align with ethical AI principles, what should be the MOST important priority?
A) Ensuring the system maximizes efficiency in identifying high-risk areas
B) Training the model on existing crime data without reviewing for bias
C) Conducting fairness audits to prevent racial or socioeconomic discrimination
D) Using a proprietary black-box model to protect competitive advantage
C) Conducting fairness audits to prevent racial or socioeconomic discrimination - Predictive policing AI must be audited to avoid biased enforcement and discriminatory impact.
Which of the following is a common method to reduce bias in AI training data?
A) Collecting more data without reviewing its quality
B) Using diverse and representative datasets
C) Ignoring potential bias to maintain AI neutrality
D) Increasing model complexity to hide biases
B) Using diverse and representative datasets - Ensuring diverse training data reduces bias and improves fairness in AI models.
What is the primary purpose of AI impact assessments?
A) To identify and mitigate potential harms associated with AI systems
B) To allow AI to operate autonomously without regulation
C) To increase AI decision speed at any cost
D) To eliminate the need for fairness audits
A) To identify and mitigate potential harms associated with AI systems - AI impact assessments help prevent risks before deployment.
What is a key challenge of AI governance in healthcare decision-making?
A) Ensuring AI models do not introduce unintended biases
B) Increasing AI automation to remove human decision-making
C) Reducing regulatory oversight for faster deployment
D) Keeping AI algorithms secret to maintain competitiveness
A) Ensuring AI models do not introduce unintended biases - AI in healthcare must be designed to provide equitable treatment to all patients.
A healthcare company is deploying a high-risk AI system for disease diagnosis. The provider of the AI system has completed a conformity assessment, but during deployment, the deployer realizes that the AI model is producing inaccurate diagnoses for certain demographic groups. Under the EU AI Act, what is the deployer’s obligation?
A) Suspend the AI system’s use, report the issue, and collaborate with the provider to retrain the model
B) Continue using the AI system and wait for additional regulations to be enforced
C) Reduce transparency in AI decision-making to minimize liability
D) Allow healthcare professionals to rely on the AI without further validation
A) Suspend the AI system’s use, report the issue, and collaborate with the provider to retrain the model - Deployers must ensure high-risk AI systems remain compliant post-deployment.
A hospital uses AI for patient risk assessment but finds that its model relies heavily on socioeconomic status as a predictor. What governance step should be taken?
A) Review and modify the AI model to prevent discrimination based on socioeconomic factors
B) Maintain the current model since it improves overall prediction accuracy
C) Remove human intervention to allow AI full control over risk assessments
D) Reduce transparency in risk scores to prevent regulatory scrutiny
A) Review and modify the AI model to prevent discrimination based on socioeconomic factors - AI in healthcare must be ethical and non-discriminatory.
Which regulation primarily governs data privacy and user rights in the European Union?
A) GDPR
B) CCPA
C) NIST RMF
D) AI Fairness Act
A) GDPR - The General Data Protection Regulation mandates data privacy controls, including AI-related data processing.
What is a significant challenge in implementing fairness constraints in AI models?
A) Balancing accuracy and fairness without introducing performance trade-offs
B) Ensuring AI operates with zero bias in all cases
C) Eliminating human oversight to improve AI efficiency
D) Keeping AI models secret to protect proprietary algorithms
B) Balancing accuracy and fairness without introducing performance trade-offs - Fairness constraints can sometimes reduce AI model accuracy.
An AI-powered financial advisory system is deployed to assist clients in making investment decisions. However, a regulatory review finds that the system lacks adequate explainability, making it difficult for clients to understand why certain recommendations are provided. What governance step should be taken?
A) Implement AI explainability mechanisms and improve user transparency
B) Maintain AI-driven investment recommendations as they optimize financial growth
C) Reduce AI transparency to protect proprietary algorithms
D) Expand AI-driven financial advising without reviewing explainability concerns
A) Implement AI explainability mechanisms and improve user transparency - Financial AI systems must be transparent and accountable.
An AI-powered mental health chatbot provides therapy suggestions. However, it occasionally makes inappropriate recommendations for users in crisis. What governance step should be taken?
A) Introduce human oversight and crisis intervention protocols
B) Maintain full AI automation for efficiency
C) Reduce transparency in chatbot responses to limit liability
D) Allow AI to continue learning without modification
A) Introduce human oversight and crisis intervention protocols - AI mental health applications must prioritize user safety and well-being.
A company developing autonomous drones wants to ensure they operate safely in populated areas. Which governance measure would BEST reduce risks?
A) Implement strict AI fail-safe mechanisms and human override controls
B) Remove human intervention to allow fully autonomous flight
C) Rely solely on historical accident data without testing in real-world conditions
D) Reduce transparency about AI flight decision-making
A) Implement strict AI fail-safe mechanisms and human override controls - Safety-critical AI systems require strong fail-safe measures.
Which principle of AI governance ensures that AI models do not produce outcomes that unfairly disadvantage specific groups?
A) Bias mitigation
B) Computational efficiency
C) Algorithmic complexity
D) Data minimization
A) Bias mitigation - Ensuring fairness in AI prevents discriminatory outcomes and promotes equity.
Which type of bias occurs when an AI system reinforces disparities present in historical data?
A) Historical bias
B) Selection bias
C) Proxy bias
D) Automation bias
A) Historical bias - This type of bias occurs when AI models learn and replicate inequalities found in past data.
Which AI category under the EU AI Act is prohibited due to unacceptable risk?
A) AI systems that manipulate human behavior or exploit vulnerabilities
B) AI models used in customer service applications
C) AI tools for business process automation
D) AI chatbots used for entertainment
A) AI systems that manipulate human behavior or exploit vulnerabilities - The EU AI Act bans AI applications that pose serious risks to individuals’ autonomy and rights.
A company is deploying an AI-driven facial recognition system at airports for identity verification. Passengers express concerns over privacy and data security. What is the MOST responsible governance approach?
A) Implement strong data protection policies and allow opt-out options
B) Expand facial recognition usage without disclosing data policies
C) Allow government agencies full access to AI-collected biometric data
D) Maintain secrecy around AI decision-making to prevent security risks
A) Implement strong data protection policies and allow opt-out options - AI-driven biometric systems must comply with privacy laws and ethical standards.
A healthcare provider deploys an AI diagnostic tool, but patients express concerns over the lack of explanation for diagnoses. What is the most appropriate governance measure?
A) Reduce AI transparency to protect proprietary algorithms
B) Ensure the AI system provides explainability features for medical staff and patients
C) Remove human oversight from AI-driven diagnosis
D) Maintain current AI decision-making to ensure efficiency
B) Ensure the AI system provides explainability features for medical staff and patients - AI in healthcare must be interpretable and trustworthy.
Which of the following is a primary threat to AI security?
A) Adversarial attacks that manipulate AI model predictions
B) AI models operating without human oversight
C) AI’s ability to process large datasets efficiently
D) The use of AI in decision automation
A) Adversarial attacks that manipulate AI model predictions - Attackers can intentionally mislead AI models by crafting deceptive inputs.
Which of the following is a major risk of using black-box AI models in high-stakes decisions?
A) Lack of explainability and transparency
B) Increased computational efficiency
C) Enhanced predictive accuracy
D) Reduced regulatory oversight
A) Lack of explainability and transparency - Black-box AI models make it difficult to assess how decisions are made.
A major ride-sharing platform deploys AI to optimize route selection. A review finds that it systematically assigns longer wait times to users in low-income areas compared to wealthier districts. What governance step should be taken?
A) Adjust AI route optimization to prevent geographic discrimination
B) Maintain AI-driven routing decisions since they improve operational efficiency
C) Reduce AI transparency in ride allocation to prevent reputational damage
D) Expand AI-driven ride optimization without reviewing fairness concerns
A) Adjust AI route optimization to prevent geographic discrimination - AI transportation services must ensure equitable access for all users.
Under the EU AI Act, what is prohibited in AI governance?
A) AI systems that exploit vulnerabilities of individuals
B) AI-driven automation in industrial manufacturing
C) AI models used in entertainment recommendations
D) AI-assisted financial modeling
C) AI systems that exploit vulnerabilities of individuals - The EU AI Act prohibits AI applications that manipulate users based on psychological or social vulnerabilities.
Which of the following is a key consideration in AI risk management?
A) Identifying potential harms and implementing mitigation strategies
B) Maximizing AI performance without considering ethical concerns
C) Ensuring AI systems operate without any regulatory oversight
D) Reducing transparency in AI decision-making to avoid legal challenges
A) Identifying potential harms and implementing mitigation strategies - AI risk management ensures that AI deployment minimizes unintended consequences.
A financial institution deploys an AI-based credit risk model. Regulators discover that the model denies loans to small business owners at a disproportionately high rate. What governance measure should be implemented?
A) Conduct a fairness audit and adjust AI decision criteria
B) Maintain AI-driven lending decisions as they optimize financial risk
C) Reduce AI transparency in loan approvals to prevent legal challenges
D) Expand AI automation in lending without modifications
A) Conduct a fairness audit and adjust AI decision criteria - AI financial systems should not reinforce systemic bias against small business owners.
A company uses AI for employee performance evaluations, and older workers receive systematically lower scores. What should be done?
A) Audit the AI system for age bias and adjust evaluation criteria
B) Maintain AI decisions since they align with company performance metrics
C) Reduce transparency in evaluations to prevent challenges
D) Automate all performance assessments without oversight
A) Audit the AI system for age bias and adjust evaluation criteria - AI workplace tools should not unfairly disadvantage specific employee demographics.
A job application AI system rejects candidates based on indirect indicators of socioeconomic status. What governance principle is MOST relevant to address this issue?
A) Fairness and bias mitigation
B) Model optimization
C) Reducing human oversight
D) Maximizing hiring efficiency
A) Fairness and bias mitigation - AI hiring systems should be designed to ensure equitable opportunities for all applicants.
Which regulation is designed to protect personal data and privacy rights in AI systems?
A) Digital Millennium Copyright Act (DMCA)
B) General Data Protection Regulation (GDPR)
C) Sarbanes-Oxley Act (SOX)
D) Basel III Accord
B) General Data Protection Regulation (GDPR) - GDPR sets strict rules on AI data processing, privacy, and consent.
A multinational corporation implements AI-driven customer service chatbots. Customers report frustration with the chatbot’s inability to understand complex issues. What governance measure should the company implement?
A) Improve natural language processing capabilities and introduce human oversight
B) Maintain the chatbot system as it reduces operational costs
C) Reduce transparency about AI limitations to manage customer expectations
D) Expand chatbot deployment without addressing concerns
A) Improve natural language processing capabilities and introduce human oversight - AI chatbots should be continuously improved to enhance user experience.
A healthcare provider implements an AI-based diagnostic tool, but patients express concerns about the lack of explanation for diagnoses. What governance measure should be prioritized?
A) Implement AI explainability features for medical staff and patients
B) Maintain AI decision-making as it optimizes diagnostic efficiency
C) Reduce transparency in AI decisions to protect proprietary information
D) Allow AI to fully automate medical diagnoses without explanation
A) Implement AI explainability features for medical staff and patients - AI explainability ensures trust and accountability in healthcare applications.
Which of the following is a primary concern in AI cybersecurity?
A) Preventing adversarial attacks that manipulate AI decision-making
B) Ensuring AI operates without any human oversight
C) Maximizing AI efficiency by reducing security constraints
D) Removing all encryption from AI models to improve transparency
A) Preventing adversarial attacks that manipulate AI decision-making - AI systems are vulnerable to attacks that exploit weaknesses in training data and algorithms.
What is the role of transparency in AI governance?
A) Ensuring AI decisions can be understood, challenged, and explained
B) Allowing AI to function autonomously without human intervention
C) Reducing regulatory compliance burdens on AI providers
D) Keeping AI decision-making confidential
C) Ensuring AI decisions can be understood, challenged, and explained - Transparency is critical for trust and accountability in AI systems.
What is the primary function of an AI ethics framework?
A) To establish guidelines for the responsible development and deployment of AI
B) To optimize AI performance without ethical considerations
C) To eliminate all human oversight in AI decision-making
D) To prevent any regulation of AI systems
A) To establish guidelines for the responsible development and deployment of AI - Ethics frameworks help ensure AI aligns with legal, social, and ethical standards.
A hospital implements an AI-powered diagnostic system that predicts patient risk for heart disease. However, the model underperforms for female patients, leading to lower diagnosis rates compared to male patients. What is the BEST approach to address this disparity?
A) Adjust the model’s threshold for all patients to ensure equal prediction rates
B) Collect and integrate additional representative training data to improve model fairness
C) Remove gender as a feature from the dataset to eliminate potential bias
D) Increase reliance on AI-generated predictions without human oversight
B) Collect and integrate additional representative training data to improve model fairness - AI fairness requires representative training data across all demographic groups.
What is the purpose of AI governance frameworks in large organizations?
A) To ensure AI development aligns with ethical, legal, and strategic business objectives
B) To eliminate the need for regulatory compliance
C) To maximize AI efficiency without oversight
D) To reduce transparency in AI decision-making
B) To ensure AI development aligns with ethical, legal, and strategic business objectives - AI governance frameworks help mitigate risks and ensure responsible AI usage.
What is differential privacy in AI governance?
A) A method to protect individual data points while training AI models
B) A strategy to increase AI efficiency by using more personal data
C) A framework that removes human oversight from AI
D) A process that ensures AI models operate with full transparency
B) A method to protect individual data points while training AI models - Differential privacy ensures user data is protected while maintaining AI accuracy.
Which of the following is an essential component of AI governance to ensure system security?
A) Data Encryption
B) AI Model Complexity
C) Reduced Human Oversight
D) Algorithmic Latency
A) Data Encryption - Securing AI models involves encrypting data to prevent unauthorized access and manipulation.
What is the primary purpose of AI model audits?
A) To assess AI fairness, accuracy, and compliance with regulations
B) To maximize AI speed and efficiency
C) To reduce human oversight in AI operations
D) To ensure AI models remain proprietary and confidential
A) To assess AI fairness, accuracy, and compliance with regulations - AI audits help prevent biases and ensure ethical decision-making.
What type of bias occurs when AI models reinforce social inequalities due to patterns in historical data?
A) Historical bias
B) Selection bias
C) Automation bias
D) Proxy bias
C) Historical bias - This occurs when AI models replicate existing inequalities present in historical data.
What is the role of an AI governance framework in an organization?
A) To establish policies for ethical AI development and deployment
B) To eliminate human oversight in AI decision-making
C) To prioritize efficiency over transparency and accountability
D) To exempt AI models from legal compliance requirements
A) To establish policies for ethical AI development and deployment - AI governance frameworks help organizations balance innovation with ethical and legal responsibilities.
What is the primary function of differential privacy in AI systems?
A) Preventing AI models from memorizing and revealing individual data points
B) Maximizing AI accuracy at the cost of user privacy
C) Ensuring AI models operate without the need for encryption
D) Allowing AI to bypass data protection regulations
B) Preventing AI models from memorizing and revealing individual data points - Differential privacy safeguards user data while maintaining model accuracy.
What is the primary function of reinforcement learning in AI?
A) Enabling AI to learn through trial and error by receiving rewards or penalties
B) Ensuring AI models are only trained with labeled datasets
C) Preventing AI from adapting its strategies over time
D) Eliminating the need for human feedback in AI training
A) Enabling AI to learn through trial and error by receiving rewards or penalties - Reinforcement learning is widely used in robotics and gaming AI.
Who is responsible for post-market monitoring of an AI system’s compliance under the EU AI Act?
A) The deployer of the AI system
B) The AI provider exclusively
C) The regulatory authority only
D) The AI system manufacturer
D) The deployer of the AI system - Deployers must continuously monitor AI system performance and report any risks or non-compliance.
A self-driving car manufacturer discovers that its AI model struggles with identifying pedestrians wearing darker clothing at night. What governance step should be taken?
A) Improve AI training data with more diverse nighttime scenarios
B) Maintain current AI performance since it works well in daylight
C) Reduce transparency in AI safety reports to prevent liability concerns
D) Expand self-driving AI deployment without additional testing
B) Improve AI training data with more diverse nighttime scenarios - AI safety must be rigorously tested in all conditions.
Which AI governance principle ensures that organizations remain responsible for AI-driven outcomes?
A) Accountability
B) Autonomy
C) Complexity
D) Speed
B) Accountability - AI systems should be developed and deployed with clear responsibilities for their outcomes.
What is the PRIMARY reason why AI systems should be designed with transparency in mind?
A) To make AI models easier to patent
B) To allow stakeholders to understand and challenge AI decisions
C) To increase computational efficiency in deep learning models
D) To eliminate the need for compliance audits
B) To allow stakeholders to understand and challenge AI decisions - Transparency ensures accountability and trust in AI systems.
A national healthcare system deploys an AI-based patient triage system to prioritize emergency cases. However, medical staff report that the AI consistently under-prioritizes patients with rare conditions, delaying critical care. Further investigation reveals that the training data lacks sufficient examples of rare diseases. What is the most responsible governance action?
A) Expand training datasets and recalibrate AI decision criteria
B) Maintain AI-driven triage decisions as they optimize hospital efficiency
C) Reduce transparency in AI triage decisions to prevent controversy
D) Automate all patient triage decisions without human oversight
B) Expand training datasets and recalibrate AI decision criteria - AI in healthcare must be equitable and accurate for all patient conditions.
A predictive policing AI disproportionately identifies crime hotspots in minority communities. What is the BEST governance response?
A) Conduct an equity audit and adjust AI decision-making criteria
B) Maintain current AI practices as they optimize crime detection
C) Reduce transparency in crime predictions to prevent challenges
D) Expand AI crime prediction to more cities without modification
A) Conduct an equity audit and adjust AI decision-making criteria - AI law enforcement tools must be designed to prevent racial and socioeconomic bias.
Which governance practice is MOST effective for mitigating risks associated with AI model drift?
A) Deploying the model without post-deployment monitoring
B) Periodically retraining the model using updated, representative data
C) Ensuring the model remains unchanged after deployment
D) Ignoring model performance degradation
B) Periodically retraining the model using updated, representative data - Model drift occurs when data patterns change over time, requiring continuous monitoring and retraining.
A company is developing an AI-powered personal assistant that schedules meetings and sends emails on behalf of users. During testing, users report that the AI sometimes discloses sensitive information without explicit permission. What governance step should the company prioritize?
A) Implement privacy-by-design principles, including user consent controls
B) Allow the AI to continue learning from interactions without restrictions
C) Reduce transparency in AI decision-making to prevent user scrutiny
D) Expand the AI’s automation capabilities without additional safeguards
A) Implement privacy-by-design principles, including user consent controls - AI systems that handle personal data must prioritize privacy and user control.
A law enforcement agency adopts an AI-powered predictive policing tool. A report finds that the system disproportionately flags low-income neighborhoods as high-crime areas, leading to increased policing in those communities. Further investigation reveals that historical crime data used for training contained inherent biases. The AI system also lacks proper oversight mechanisms. What governance action should be taken?
A) Conduct a fairness audit, retrain the AI model, and establish oversight protocols
B) Maintain AI-driven crime prediction since it improves policing efficiency
C) Reduce AI transparency in law enforcement to avoid public controversy
D) Expand AI-driven policing tools without addressing bias concerns
B) Conduct a fairness audit, retrain the AI model, and establish oversight protocols - AI in law enforcement must be unbiased and transparent.
An AI-powered mortgage approval system is found to approve loans at a significantly lower rate for self-employed applicants. What governance action should be taken?
A) Conduct an audit to identify bias and adjust approval criteria
B) Maintain the system since it aligns with historical lending trends
C) Reduce transparency in loan approvals to avoid regulatory challenges
D) Remove all human oversight to ensure AI neutrality
A) Conduct an audit to identify bias and adjust approval criteria - AI in finance must ensure fair and unbiased lending decisions.
A social media platform’s AI content moderation system is flagged under EU AI Act compliance reviews for suppressing certain political viewpoints. What governance measure should be taken?
A) Ensure AI content moderation algorithms do not disproportionately impact specific viewpoints
B) Maintain AI-driven moderation as it improves user engagement
C) Reduce AI transparency in content moderation to prevent regulatory intervention
D) Expand AI-driven content filtering without fairness reviews
A) Ensure AI content moderation algorithms do not disproportionately impact specific viewpoints - AI must adhere to fairness and freedom of expression protections.
An AI-powered investment tool advises clients on portfolio management. Regulators discover that the AI promotes riskier investments to lower-income users. What governance step should be taken?
A) Adjust AI decision-making to ensure fair investment recommendations
B) Maintain current AI behavior as it optimizes financial returns
C) Reduce transparency in financial AI decision-making to prevent regulatory intervention
D) Expand AI automation to reduce reliance on human financial advisors
A) Adjust AI decision-making to ensure fair investment recommendations - AI in financial services must provide ethical and unbiased investment guidance.
A healthcare AI system is found to diagnose a particular disease less accurately for minority populations due to underrepresentation in training data. What is the best governance response?
A) Expand training datasets and retrain AI models to ensure equitable outcomes
B) Maintain the current AI model since it performs well for the majority population
C) Reduce transparency about AI decision-making to avoid controversy
D) Automate all medical diagnoses without additional human oversight
A) Expand training datasets and retrain AI models to ensure equitable outcomes - AI in healthcare must provide fair treatment for all patient demographics.
An AI-driven content moderation system is found to disproportionately flag posts from non-native English speakers. What governance measure should be implemented?
A) Conduct a fairness audit and adjust moderation criteria
B) Maintain current AI filtering as it improves platform safety
C) Reduce AI transparency to prevent user complaints
D) Expand AI-driven content moderation without reviewing bias concerns
A) Conduct a fairness audit and adjust moderation criteria - AI moderation systems must ensure fairness across different language and cultural groups.
A healthcare AI system prioritizes treatment recommendations based on patient age, leading to older patients receiving lower priority for critical care. What governance step should be prioritized?
A) Review and adjust AI decision criteria to ensure ethical fairness
B) Maintain AI-driven recommendations since they improve efficiency
C) Reduce AI transparency in medical decision-making
D) Automate all treatment prioritization without human oversight
C) Review and adjust AI decision criteria to ensure ethical fairness - AI in healthcare should ensure equitable treatment for all patient demographics.
A hospital uses AI for diagnosing rare diseases but finds that it is significantly less accurate for certain demographic groups. What governance step should be taken?
A) Expand training datasets and recalibrate AI predictions
B) Maintain AI diagnostic decisions as they improve efficiency
C) Reduce transparency in AI decision-making to avoid public backlash
D) Automate all diagnoses without human involvement
B) Expand training datasets and recalibrate AI predictions - AI in healthcare must be trained on diverse patient groups to ensure fairness.
A company uses an AI model to optimize employee work schedules. Employees complain that the AI assigns shifts unfairly, favoring certain workers. What governance step should the company take?
A) Conduct an audit and adjust the AI algorithm for fairness
B) Keep the AI system unchanged since it improves operational efficiency
C) Reduce transparency in scheduling decisions to prevent disputes
D) Allow AI to automate all scheduling without human oversight
A) Conduct an audit and adjust the AI algorithm for fairness - AI workforce management should prioritize fairness and avoid bias.
A company uses AI to personalize online advertising but discovers that certain groups receive fewer job-related ads. What governance measure should be taken?
A) Implement fairness constraints to ensure balanced ad distribution
B) Maintain current AI decisions since they optimize engagement
C) Reduce AI transparency in ad targeting to prevent scrutiny
D) Expand AI-based advertising without adjusting bias concerns
A) Implement fairness constraints to ensure balanced ad distribution - AI-driven advertising should not reinforce social inequalities.
What is a primary risk of black-box AI models in high-stakes decision-making?
A) Lack of transparency and interpretability
B) Increased computational efficiency
C) Reduced bias in decision-making
D) Greater regulatory compliance
A) Lack of transparency and interpretability - Black-box AI models make it difficult to ensure fairness and accountability.
A company develops an AI tool for talent acquisition. The AI systematically rejects applicants from non-traditional educational backgrounds. What governance step should the company take?
A) Conduct a fairness audit and modify AI decision criteria
B) Maintain AI automation as it optimizes hiring efficiency
C) Reduce transparency in AI recruitment decisions to prevent legal challenges
D) Expand AI decision-making without modifying hiring factors
A) Conduct a fairness audit and modify AI decision criteria - AI recruitment should not reinforce exclusionary hiring practices.
What is the main goal of AI fairness in governance frameworks?
A) To prevent AI from producing biased or discriminatory outcomes
B) To maximize AI efficiency without ethical considerations
C) To ensure AI models remain proprietary and confidential
D) To remove human oversight from AI decision-making
B) To prevent AI from producing biased or discriminatory outcomes - AI fairness ensures equal treatment in AI-driven decisions.
A healthcare AI model predicts patient readmission risks but disproportionately flags individuals from low-income backgrounds. What governance step should be prioritized?
A) Audit the training data for bias and implement fairness adjustments
B) Maintain the AI model since it improves hospital efficiency
C) Reduce transparency in AI decision-making to prevent challenges
D) Automate all healthcare decisions without human intervention
A) Audit the training data for bias and implement fairness adjustments - AI in healthcare should ensure equitable outcomes across patient demographics.
A social media platform uses AI to filter harmful content, but users claim it disproportionately censors certain viewpoints. What is the MOST ethical response?
A) Conduct fairness audits and adjust moderation policies for balanced content filtering
B) Maintain current AI decisions to maximize efficiency
C) Remove transparency in content moderation to prevent user scrutiny
D) Expand AI censorship capabilities without oversight
A) Conduct fairness audits and adjust moderation policies for balanced content filtering - AI moderation must balance fairness with harm prevention.
A law enforcement agency adopts AI facial recognition technology for public surveillance. Investigations reveal that the AI system disproportionately misidentifies individuals from minority groups. However, officials refuse to disclose the AI training process. What governance action should be taken?
A) Implement transparency requirements and conduct fairness audits on the AI system
B) Maintain AI-driven surveillance since it improves public safety
C) Reduce AI transparency to avoid public scrutiny
D) Expand AI-driven surveillance without reviewing accuracy concerns
A) Implement transparency requirements and conduct fairness audits on the AI system - AI in law enforcement must be transparent, fair, and accountable.
What is a primary method for mitigating bias in AI systems?
A) Using diverse and representative training datasets
B) Eliminating regulatory oversight to encourage rapid AI deployment
C) Ensuring AI operates autonomously without human intervention
D) Keeping AI decision-making criteria confidential
B) Using diverse and representative training datasets - Training AI on balanced datasets reduces bias in decision-making.
A hospital deploys an AI-powered patient monitoring system that uses deep learning to detect early signs of deterioration. However, the system’s accuracy is significantly lower for patients with rare conditions due to insufficient training data. What governance measure should be prioritized?
A) Expand training datasets with more diverse patient data and retrain the AI model
B) Maintain AI-driven monitoring as it improves efficiency for most patients
C) Reduce transparency in AI decision-making to avoid patient concerns
D) Automate all monitoring decisions without human review
C) Expand training datasets with more diverse patient data and retrain the AI model - AI models in healthcare must be trained on diverse datasets to ensure accuracy.
What is AI model governance?
A) A framework ensuring AI models comply with regulatory, ethical, and operational standards
B) A system designed to eliminate human intervention in AI decision-making
C) A strategy to prioritize AI efficiency over fairness
D) A process that keeps AI decision-making confidential
A) A framework ensuring AI models comply with regulatory, ethical, and operational standards - AI governance ensures responsible development and deployment.
How can organizations ensure fairness in AI decision-making processes?
A) By regularly auditing AI models for bias and unintended discrimination
B) By allowing AI models to make independent decisions without human oversight
C) By keeping AI model parameters secret to avoid external influence
D) By prioritizing efficiency over fairness in AI deployment
A) By regularly auditing AI models for bias and unintended discrimination - Continuous monitoring ensures AI fairness and compliance.
What is the primary purpose of AI impact assessments?
A) To evaluate and mitigate potential ethical, legal, and social risks
B) To remove all regulatory constraints from AI deployment
C) To make AI models more complex and difficult to understand
D) To prevent human oversight in AI decision-making
A) To evaluate and mitigate potential ethical, legal, and social risks - AI impact assessments help organizations identify potential negative outcomes before deployment.
An AI-powered investment tool is found to suggest riskier investment strategies to low-income users compared to wealthier clients. What governance step should be taken?
A) Maintain AI investment advice as it maximizes returns
B) Conduct a fairness audit and adjust recommendation criteria
C) Reduce AI transparency in financial recommendations to avoid scrutiny
D) Expand AI financial advice services without modification
D) Conduct a fairness audit and adjust recommendation criteria - AI financial services should be fair and unbiased.
A healthcare AI system under EU AI Act compliance is found to prioritize treatment based on insurance status rather than medical urgency. What governance measure should be implemented?
A) Adjust AI decision criteria to ensure equitable healthcare access
B) Maintain AI-driven treatment prioritization for efficiency
C) Reduce transparency in AI healthcare decisions to prevent complaints
D) Expand AI-driven medical decision-making without reviewing fairness
A) Adjust AI decision criteria to ensure equitable healthcare access - AI in healthcare must align with ethical principles and fairness requirements.
Which of the following is a defining characteristic of deep learning models?
A) The use of multiple layers in neural networks to extract complex patterns
B) The ability to function without training data
C) The reliance on fixed decision rules instead of adaptive learning
D) The requirement that all AI decisions be explainable
A) The use of multiple layers in neural networks to extract complex patterns - Deep learning models utilize neural networks to process and learn from large datasets.
A social media platform implements an AI moderation system that disproportionately removes content from minority users. The platform claims that the AI was trained on large-scale data and is neutral. What governance strategy should the company adopt?
A) Conduct an external audit and introduce fairness metrics to prevent discriminatory moderation
B) Rely solely on AI decision-making and remove human oversight to reduce bias
C) Reduce transparency to avoid public scrutiny
D) Expand AI enforcement across all content categories to appear neutral
A) Conduct an external audit and introduce fairness metrics to prevent discriminatory moderation - AI content moderation must be fair and inclusive to avoid bias.
Which of the following is a key risk associated with black-box AI models?
A) They improve explainability and regulatory compliance
B) Their decision-making processes are difficult to interpret
C) They always produce unbiased and ethical outcomes
D) They require less data than transparent models
B) Their decision-making processes are difficult to interpret - Black-box AI models can create accountability issues due to lack of transparency.
A hospital uses an AI-driven triage system to prioritize emergency room patients. It is discovered that patients from non-English-speaking backgrounds are often assigned lower priority. What governance measure should be implemented?
A) Conduct a fairness audit and improve linguistic diversity in training data
B) Maintain current triage decisions as they optimize patient flow
C) Reduce transparency in AI-driven prioritization to avoid legal scrutiny
D) Automate all patient triage decisions to remove human bias
A) Conduct a fairness audit and improve linguistic diversity in training data - AI in healthcare must ensure fair and equitable treatment for all patients.
A national security agency deploys AI-powered surveillance but faces criticism for potential privacy violations. What is the MOST ethical governance approach?
A) Implement strict oversight and privacy-preserving AI techniques
B) Expand AI surveillance without public disclosure to enhance security
C) Remove regulatory oversight to improve AI efficiency
D) Limit transparency to prevent public pushback
A) Implement strict oversight and privacy-preserving AI techniques - AI surveillance must balance security with ethical privacy protections.
Which of the following is an example of an AI risk that requires mitigation strategies?
A) AI systems that generate biased outcomes due to skewed training data
B) AI models that always make perfect predictions
C) AI systems that require minimal data for decision-making
D) AI models that operate without any human oversight
C) AI systems that generate biased outcomes due to skewed training data - AI bias can lead to unfair or unethical outcomes.
How does the ‘Govern’ function of the NIST AI Risk Management Framework (RMF) support AI governance?
A) Establishing accountability structures and oversight for AI risk management
B) Ensuring AI operates without human intervention
C) Maximizing computational efficiency at the cost of ethical considerations
D) Removing the need for AI compliance checks
C) Establishing accountability structures and oversight for AI risk management - Governance is key to ensuring responsible AI deployment.
An AI-powered content moderation system flags a higher percentage of posts from minority users. What governance action should be taken?
A) Conduct an impact assessment and retrain the AI model to reduce bias
B) Maintain the current model since it reduces overall harmful content
C) Reduce AI transparency to avoid public backlash
D) Expand AI content moderation without adjusting bias factors
A) Conduct an impact assessment and retrain the AI model to reduce bias - AI content moderation must be equitable and non-discriminatory.
A multinational company implements an AI-driven fraud detection system. A regulatory investigation finds that the AI disproportionately flags transactions from specific ethnic communities for additional scrutiny. What is the most appropriate governance action?
A) Conduct a fairness audit and recalibrate fraud detection criteria
B) Maintain AI-driven fraud detection since it optimizes financial security
C) Reduce AI transparency to prevent customer complaints
D) Expand AI fraud detection without reviewing bias concerns
C) Conduct a fairness audit and recalibrate fraud detection criteria - AI in financial security should not reinforce systemic discrimination.
Which AI principle ensures that AI decisions are traceable and understandable?
A) Explainability
B) Efficiency
C) Latency Reduction
D) Model Complexity
A) Explainability - AI systems should provide insights into how decisions are made, enhancing transparency and trust.
Which of the following BEST describes algorithmic accountability?
A) Ensuring AI developers take responsibility for system outcomes
B) Reducing human oversight to improve AI efficiency
C) Allowing AI models to operate without legal restrictions
D) Keeping AI decision-making opaque to maintain competitiveness
A) Ensuring AI developers take responsibility for system outcomes - Accountability in AI governance ensures responsible AI development and deployment.
An AI-driven content moderation system is found to disproportionately flag posts from non-English-speaking users. What governance measure should be implemented?
A) Conduct a fairness audit and adjust moderation criteria
B) Maintain current AI filtering as it improves platform safety
C) Reduce AI transparency to avoid scrutiny
D) Expand AI moderation without reviewing for language bias
A) Conduct a fairness audit and adjust moderation criteria - AI moderation must ensure fairness across linguistic and cultural groups.
A hospital implements an AI-based medical diagnosis tool. However, doctors report that the AI system does not provide justifications for its diagnostic decisions, making it difficult to trust the recommendations. What governance step should be taken?
A) Introduce explainability mechanisms that allow medical professionals to understand AI-driven conclusions
B) Maintain AI-driven diagnoses as they optimize hospital efficiency
C) Reduce AI transparency to prevent legal liability
D) Automate all diagnostic decisions without requiring physician review
C) Introduce explainability mechanisms that allow medical professionals to understand AI-driven conclusions - AI in healthcare must be transparent and interpretable.
Which element of the NIST AI Risk Management Framework (RMF) ensures that organizations continuously monitor AI risks post-deployment?
A) The ‘Manage’ function, which integrates risk mitigation into ongoing operations
B) The ‘Map’ function, which primarily focuses on initial risk identification
C) The ‘Measure’ function, which focuses only on model accuracy
D) The ‘Govern’ function, which applies solely to pre-deployment risk assessments
B) The ‘Manage’ function, which integrates risk mitigation into ongoing operations - AI risk management must be continuous and adaptive.
A ride-sharing company deploys an AI-powered pricing model. A review finds that fares are disproportionately higher in low-income neighborhoods compared to wealthier districts, leading to affordability concerns. What governance action should the company take?
A) Adjust AI pricing criteria to prevent geographic discrimination
B) Maintain AI-driven pricing as it follows market demand
C) Reduce AI transparency in fare adjustments to avoid user backlash
D) Expand AI-driven surge pricing without addressing fairness issues
D) Adjust AI pricing criteria to prevent geographic discrimination - AI pricing models should be equitable across different regions.
What is algorithmic bias in AI?
A) The tendency of an AI system to produce outcomes that systematically disadvantage certain groups
B) The ability of AI to detect human biases
C) The use of AI to remove bias from human decisions
D) A technique used to make AI more efficient
A) The tendency of an AI system to produce outcomes that systematically disadvantage certain groups - Algorithmic bias occurs when AI models reflect and perpetuate existing biases in training data.
Which of the following would be considered a ‘black box’ AI model?
A) A model whose decisions can be easily traced and explained
B) A simple linear regression model with transparent weights
C) A deep neural network with complex decision-making that lacks interpretability
D) A decision tree model with clear branching logic
C) A deep neural network with complex decision-making that lacks interpretability - ‘Black box’ AI refers to models where decision processes are opaque and difficult to explain.
A predictive policing AI tool disproportionately flags low-income neighborhoods as high-risk areas. What governance response should be prioritized?
A) Conduct an equity audit and recalibrate AI policing criteria
B) Maintain AI-driven policing decisions as they optimize crime detection
C) Reduce transparency in AI policing to prevent community concerns
D) Expand AI crime prediction tools without modification
A) Conduct an equity audit and recalibrate AI policing criteria - AI in law enforcement must be fair and non-discriminatory.
A company deploys an AI-powered fraud detection system, but a review finds it disproportionately flags transactions from small businesses. What is the best governance response?
A) Conduct a bias audit and recalibrate the fraud detection model
B) Maintain AI fraud detection as it improves financial security
C) Reduce transparency in fraud detection criteria to prevent scrutiny
D) Expand AI-driven fraud detection without reviewing bias concerns
A) Conduct a bias audit and recalibrate the fraud detection model - AI fraud detection must be fair and unbiased in assessing transactions.
What is an adversarial attack in AI security?
A) A method of manipulating AI models by introducing deceptive inputs
B) A process of improving AI accuracy through external interventions
C) A technique used to remove bias from AI systems
D) A method of increasing AI efficiency without security concerns
A) A method of manipulating AI models by introducing deceptive inputs - Adversarial attacks trick AI into making incorrect predictions.
What is machine learning in the context of AI?
A) A subset of AI that enables systems to learn and improve from data
B) A process where AI models operate without any training data
C) A method used to hardcode all AI decision-making rules
D) A form of AI that does not require updates or retraining
A) A subset of AI that enables systems to learn and improve from data - Machine learning allows AI to recognize patterns and make predictions without explicit programming.
What is the primary role of explainability in AI risk management?
A) Ensuring that AI decisions are interpretable and accountable to stakeholders
B) Maximizing AI computational efficiency
C) Reducing regulatory compliance burdens
D) Keeping AI decision-making confidential
B) Ensuring that AI decisions are interpretable and accountable to stakeholders - Explainability improves trust and compliance.
What is a common challenge in AI bias mitigation?
A) Ensuring AI models do not unintentionally favor one group over another
B) Removing fairness constraints to optimize AI performance
C) Reducing transparency in AI decision-making to prevent scrutiny
D) Eliminating regulatory oversight on AI fairness
A) Ensuring AI models do not unintentionally favor one group over another - AI fairness requires continuous monitoring and intervention.
Which of the following best defines algorithmic fairness in AI governance?
A) The principle that AI should not produce discriminatory outcomes
B) The idea that AI should always prioritize efficiency over fairness
C) The requirement that AI models be completely transparent to the public
D) The belief that AI bias is inevitable and does not require mitigation
A) The principle that AI should not produce discriminatory outcomes - Algorithmic fairness ensures equitable treatment across all user groups.
An AI-powered investment platform adjusts trading strategies based on user data. Regulators raise concerns about potential market manipulation. What is the MOST responsible action?
A) Conduct an AI ethics audit and implement safeguards against manipulation
B) Maintain AI decision-making secrecy to protect proprietary algorithms
C) Increase automation to further enhance trading performance
D) Reduce transparency in AI trading mechanisms
A) Conduct an AI ethics audit and implement safeguards against manipulation - AI-driven finance must be regulated to prevent unethical market influence.
A city government deploys an AI-based traffic control system. A study finds that congestion has worsened in low-income neighborhoods. What governance step should be prioritized?
A) Adjust AI traffic optimization algorithms to ensure equitable outcomes
B) Maintain current AI traffic patterns as they improve overall efficiency
C) Reduce transparency in AI traffic control decisions to prevent opposition
D) Expand AI-driven traffic control without modification
A) Adjust AI traffic optimization algorithms to ensure equitable outcomes - AI in public infrastructure should provide benefits for all communities.
A hospital deploys an AI-driven diagnostic system that frequently misdiagnoses elderly patients. What governance action is MOST appropriate?
A) Improve training data diversity and adjust model fairness constraints
B) Maintain the AI system since it works well for younger patients
C) Reduce AI transparency to avoid legal liability
D) Allow AI to fully automate medical decisions without human intervention
A) Improve training data diversity and adjust model fairness constraints - AI healthcare tools should be designed to work effectively for all patient groups.
A government agency implements AI for social welfare eligibility decisions. Reports indicate that applicants from specific communities face higher rejection rates. What is the most responsible response?
A) Conduct a fairness audit and introduce bias mitigation strategies
B) Maintain the current model since it streamlines application processing
C) Reduce transparency in eligibility decisions to prevent scrutiny
D) Expand AI decision-making without reviewing bias concerns
A) Conduct a fairness audit and introduce bias mitigation strategies - AI in public services must ensure equitable and unbiased decision-making.
How does the EU AI Act categorize AI systems that present an unacceptable risk?
A) They are banned from use within the EU unless specific exemptions apply
B) They must undergo periodic regulatory assessment but remain in operation
C) They are allowed with transparency obligations but require human oversight
D) They are classified as low-risk systems and face minimal restrictions
A) They are banned from use within the EU unless specific exemptions apply - The EU AI Act prohibits AI applications deemed excessively harmful.
A social media platform’s AI algorithm prioritizes highly emotional content because it increases engagement. What is the BEST governance response?
A) Adjust the algorithm to prioritize accuracy and credibility over engagement
B) Maintain AI-driven ranking for maximum revenue
C) Reduce transparency in content moderation to prevent criticism
D) Expand AI-driven content recommendations without modification
B) Adjust the algorithm to prioritize accuracy and credibility over engagement - AI should prevent the spread of misinformation.
Which OECD HUDERAF principle directly addresses the need for AI systems to function reliably under diverse and adverse conditions?
A) Robustness, which ensures AI remains resilient against errors and adversarial attacks
B) Diversity, which promotes demographic representation in training data
C) Universality, which mandates AI accessibility for all users
D) Explainability, which focuses on interpretability rather than system reliability
C) Robustness, which ensures AI remains resilient against errors and adversarial attacks - AI robustness protects against instability and malicious interference.
A university deploys AI-driven admissions screening. A study finds that students from rural areas are less likely to be admitted. What is the most ethical governance action?
A) Conduct an equity audit and adjust AI decision criteria
B) Maintain AI admissions decisions as they improve efficiency
C) Reduce AI transparency to prevent challenges from rejected applicants
D) Automate all admissions decisions without human oversight
A) Conduct an equity audit and adjust AI decision criteria - AI in education should ensure fair access for all applicants.
An AI-powered hiring platform is found to consistently favor male candidates over female applicants. What governance measure should be implemented?
A) Conduct an equity audit and adjust AI decision criteria
B) Maintain current AI hiring decisions as they align with past trends
C) Reduce transparency in AI-driven hiring to avoid legal challenges
D) Expand AI automation in hiring without modifications
A) Conduct an equity audit and adjust AI decision criteria - AI recruitment systems should ensure fair and equal hiring opportunities.
What is a significant limitation of differential privacy as an AI governance tool?
A) It introduces a trade-off between privacy protection and model accuracy
B) It eliminates all potential risks associated with data leaks
C) It ensures AI models remain entirely explainable without any bias
D) It allows AI systems to operate without the need for regulatory compliance
B) It introduces a trade-off between privacy protection and model accuracy - Differential privacy techniques can reduce model precision while safeguarding individual data.
A financial services company deploys an AI-driven loan approval system developed by a third-party provider. After six months, the deployer identifies that the system consistently rejects loan applications from minority groups at a disproportionate rate. According to the EU AI Act, what is the deployer’s responsibility in this case?
A) Report the issue to regulatory authorities and work with the provider to implement corrective measures
B) Continue using the AI system while awaiting further clarification from the provider
C) Disable transparency mechanisms to avoid scrutiny
D) Expand AI deployment without modifying risk assessment criteria
D) Report the issue to regulatory authorities and work with the provider to implement corrective measures - Deployers must monitor and mitigate post-market AI risks.
What is a key concern with black-box AI models in high-stakes decision-making?
A) They may lack explainability and make it difficult to understand AI decisions
B) They require less computational power compared to transparent models
C) They eliminate all bias in decision-making
D) They are more expensive to implement
A) They may lack explainability and make it difficult to understand AI decisions - Black-box AI systems pose accountability risks in critical applications.
A university uses an AI-based admissions system that prioritizes students from affluent backgrounds. What governance measure should be implemented?
A) Conduct an impact assessment and refine AI ranking criteria
B) Maintain AI admissions rankings as they align with historical enrollment trends
C) Reduce transparency in AI decision-making to prevent student complaints
D) Automate all admissions decisions without reviewing fairness factors
A) Conduct an impact assessment and refine AI ranking criteria - AI in education should ensure fair and equal access to opportunities.
An AI-powered hiring system is found to systematically reject older applicants. What governance measure should be taken?
A) Conduct an age bias audit and retrain the AI model
B) Maintain AI decisions as they reflect past hiring success
C) Reduce transparency in hiring decisions to avoid complaints
D) Automate hiring without reviewing AI recommendations
A) Conduct an age bias audit and retrain the AI model - AI in hiring must comply with anti-discrimination regulations.
An AI-powered loan approval system consistently approves fewer applications from minority communities despite similar financial qualifications. What is the most responsible governance action?
A) Conduct a bias audit and modify the AI model to ensure fairness
B) Maintain the system as is since it optimizes financial risk assessment
C) Remove transparency in AI decision-making to avoid regulatory issues
D) Increase AI autonomy to eliminate human involvement
A) Conduct a bias audit and modify the AI model to ensure fairness - Financial AI systems must be free from systemic discrimination.
Which governance principle ensures that AI systems do not operate with unchecked power?
A) Human oversight and intervention
B) Algorithmic complexity
C) AI efficiency optimization
D) Data minimization
A) Human oversight and intervention - AI governance must include human checks to prevent harmful automated decisions.
Which global framework provides guidelines for AI risk management and compliance?
A) NIST AI Risk Management Framework (RMF)
B) HIPAA
C) Fair Labor Standards Act
D) The Paris Agreement
A) NIST AI Risk Management Framework (RMF) - The NIST RMF helps organizations identify, assess, and mitigate AI risks.
A law enforcement agency uses an AI-powered facial recognition system that incorrectly identifies individuals from certain ethnic groups more frequently. What is the most ethical response?
A) Increase police reliance on AI since it still reduces crime overall
B) Conduct a bias audit and retrain the model with diverse datasets
C) Reduce transparency in AI decision-making to avoid public criticism
D) Maintain the current system because biometric AI is inherently neutral
B) Conduct a bias audit and retrain the model with diverse datasets - AI law enforcement tools must be free from bias to ensure fair and just outcomes.
Which of the following AI principles ensures users are informed about AI’s role in decision-making?
A) Transparency
B) Complexity
C) Algorithmic secrecy
D) Full automation
A) Transparency - Transparency in AI governance ensures users understand AI decision-making processes and impacts.
What is the main ethical concern regarding AI-powered predictive policing?
A) AI models may reinforce systemic bias and disproportionately target certain communities
B) AI-driven policing always improves law enforcement effectiveness
C) AI does not require regulatory oversight in policing applications
D) AI-based crime prediction is infallible
A) AI models may reinforce systemic bias and disproportionately target certain communities - Predictive policing AI must be monitored to prevent discriminatory enforcement.
A self-driving car company is designing an AI decision-making model for accident scenarios. What is the MOST ethical consideration?
A) Ensuring decisions are made based on legal liability minimization
B) Prioritizing passenger safety while considering external risks
C) Reducing transparency about AI decision logic to avoid controversy
D) Allowing AI to fully automate accident response without oversight
B) Prioritizing passenger safety while considering external risks - Autonomous vehicle AI must be designed with ethical safety considerations.
Which AI governance principle is MOST critical when developing an AI system that predicts job candidate success?
A) Cost efficiency
B) Automation of all decision-making
C) Bias mitigation and fairness
D) Model complexity
C) Bias mitigation and fairness - AI hiring tools must ensure fair treatment and avoid discriminatory bias.
An AI-powered mental health chatbot provides automated therapy suggestions. However, users report that it sometimes responds inappropriately to crisis situations. What should be the company’s highest priority?
A) Implement human intervention protocols for high-risk cases
B) Allow the chatbot to continue without modifications to ensure full automation
C) Reduce transparency in chatbot decision-making to prevent liability
D) Expand the chatbot’s reach without addressing the crisis response issue
A) Implement human intervention protocols for high-risk cases - AI in mental health applications must include safeguards to prevent harm.
Which of the following represents a potential pitfall of over-reliance on AI explainability techniques in high-stakes decision-making?
A) The possibility of misleading explanations that do not accurately reflect model behavior
B) The elimination of all bias from AI decision-making
C) The ability to fully automate regulatory compliance processes
D) The reduction of human oversight in AI governance
D) The possibility of misleading explanations that do not accurately reflect model behavior - Explainability techniques can sometimes oversimplify or distort model decision logic.
A government agency uses AI to detect welfare fraud. However, it disproportionately flags single-parent households. What is the BEST response?
A) Audit and recalibrate the AI system for fairness
B) Maintain current fraud detection models to ensure efficiency
C) Reduce transparency in fraud detection to prevent challenges
D) Expand AI surveillance without modifying risk factors
A) Audit and recalibrate the AI system for fairness - AI in public services must be unbiased and fair to all users.
Which of the following is a requirement for AI transparency under the EU AI Act?
A) AI systems must provide meaningful explanations of automated decisions to users
B) AI models must be kept confidential to protect intellectual property
C) AI transparency only applies to high-risk applications
D) Organizations must eliminate explainability to improve AI efficiency
A) AI systems must provide meaningful explanations of automated decisions to users - The EU AI Act mandates transparency to ensure accountability.
A major e-commerce platform deploys an AI recommendation system for products. A customer advocacy group finds that certain demographic groups receive significantly different recommendations, but the company refuses to disclose how the AI model operates. What governance measure should be implemented?
A) Increase transparency and provide users with explanations for AI-generated recommendations
B) Maintain AI-driven recommendations as they improve user engagement
C) Reduce AI transparency to protect competitive advantage
D) Expand AI-driven recommendations without addressing transparency concerns
D) Increase transparency and provide users with explanations for AI-generated recommendations - AI recommendation systems should be explainable and unbiased.
Which category of AI systems is considered ‘high-risk’ under the EU AI Act?
A) AI systems used in critical infrastructure, law enforcement, and healthcare
B) AI models used for entertainment and gaming applications
C) AI chatbots designed for customer service
D) AI models deployed solely for research purposes
A) AI systems used in critical infrastructure, law enforcement, and healthcare - High-risk AI applications require stricter regulatory oversight under the EU AI Act.
Which of the following is a primary goal of AI fairness?
A) Ensuring AI systems produce equitable outcomes across different demographic groups
B) Maximizing AI efficiency at the cost of fairness
C) Removing all human oversight from AI decision-making
D) Keeping AI models confidential to protect proprietary interests
A) Ensuring AI systems produce equitable outcomes across different demographic groups - Fairness in AI prevents bias and ensures ethical decision-making.
A hospital deploys an AI-based diagnostic tool that provides highly accurate results but is not interpretable by medical professionals. Doctors express concern over using an AI system they do not understand. What should the hospital prioritize?
A) Continue AI deployment without explanation to improve efficiency
B) Increase explainability and interpretability of AI predictions before widespread use
C) Automate decision-making and remove doctors from the process
D) Reduce the AI system’s accuracy to make it more interpretable
B) Increase explainability and interpretability of AI predictions before widespread use - Transparency in AI decision-making is critical, especially in healthcare.
A financial institution deploys an AI-driven credit approval system. Customers who receive rejections request an explanation, but the bank cannot provide one because the AI model is a black-box system. What governance action should be prioritized?
A) Implement explainability techniques and ensure AI decisions are interpretable
B) Maintain AI-driven credit decisions as they optimize risk assessment
C) Reduce AI transparency in loan approvals to prevent scrutiny
D) Expand AI lending automation without addressing explainability concerns
B) Implement explainability techniques and ensure AI decisions are interpretable - Financial AI systems must provide understandable decision rationales.
A financial institution deploys an AI-powered loan approval system. A review finds that applicants from low-income neighborhoods receive higher rejection rates than equally qualified applicants from wealthier areas. What is the most appropriate governance response?
A) Conduct a fairness audit and adjust the AI model to ensure equitable lending practices
B) Maintain AI-driven loan approvals since they optimize financial risk assessment
C) Reduce AI transparency in loan decisions to prevent regulatory scrutiny
D) Expand AI-driven lending without modifying fairness factors
A) Conduct a fairness audit and adjust the AI model to ensure equitable lending practices - AI in finance should ensure fairness across demographic groups.
A smart city deploys AI traffic control. Data analysis shows that traffic congestion has worsened in lower-income neighborhoods. What governance step should be prioritized?
A) Adjust AI traffic distribution algorithms to ensure equitable outcomes
B) Maintain current AI traffic control since it optimizes overall efficiency
C) Reduce AI transparency to prevent community opposition
D) Expand AI-driven traffic control without addressing concerns
A) Adjust AI traffic distribution algorithms to ensure equitable outcomes - AI in public infrastructure should benefit all communities equally.
A hospital network implements an AI-powered diagnostic system to assist physicians. While the system performs well for common diseases, an internal study finds that it misdiagnoses rare conditions at a significantly higher rate due to limited training data. In addition, physicians report difficulty interpreting AI recommendations due to the model’s black-box nature. How should the hospital respond?
A) Expand training datasets and implement explainability mechanisms for AI recommendations
B) Maintain AI-driven diagnostics as they improve efficiency for the majority of cases
C) Reduce AI transparency to prevent patient distrust
D) Automate all diagnoses without physician oversight
A) Expand training datasets and implement explainability mechanisms for AI recommendations - AI in healthcare should be accurate, explainable, and equitable.
A healthcare provider deploys an AI diagnostic tool that outperforms human doctors in some cases but lacks explainability. Which ethical principle is MOST at risk?
A) Autonomy
B) Privacy
C) Transparency
D) Beneficence
C) Transparency - Lack of explainability in high-stakes AI applications can reduce trust and accountability.
An AI recruitment tool prioritizes applicants from certain universities, reducing diversity. What is the BEST governance step?
A) Conduct a bias audit and adjust AI ranking criteria
B) Maintain the AI system since it follows historical hiring patterns
C) Reduce transparency in AI selection criteria to prevent complaints
D) Expand AI automation in hiring without adjusting bias factors
A) Conduct a bias audit and adjust AI ranking criteria - AI recruitment must ensure fair opportunities for all applicants.
Which type of AI is designed to function at or beyond human cognitive capabilities across all domains?
A) Artificial General Intelligence (AGI)
B) Narrow AI (ANI)
C) Weak AI
D) Expert Systems
D) Artificial General Intelligence (AGI) - AGI refers to AI systems capable of reasoning and problem-solving across multiple tasks, similar to human cognition.
What is the main goal of AI model documentation?
A) To provide transparency, accountability, and traceability in AI decision-making
B) To ensure AI operates without human oversight
C) To make AI models proprietary and confidential
D) To prevent external audits of AI systems
A) To provide transparency, accountability, and traceability in AI decision-making - Proper documentation ensures AI systems remain understandable and auditable.
Which AI governance principle ensures that organizations take responsibility for AI-related risks and harms?
A) Accountability
B) Complexity
C) Scalability
D) Data obfuscation
D) Accountability - AI developers and deployers must take responsibility for AI decisions and their consequences.
Which of the following is a core requirement for AI explainability?
A) The AI system must be fully automated without human oversight
B) Users must be able to understand how decisions are made
C) AI models should prioritize efficiency over interpretability
D) AI developers should keep decision-making processes confidential
B) Users must be able to understand how decisions are made - AI explainability ensures accountability and trust in automated decision-making.
An AI hiring system is trained to screen job applications but is found to favor applicants from certain universities due to biases in the training data. The system was trained using historical hiring patterns that reflect existing disparities. What governance step should be prioritized?
A) Conduct a bias audit and retrain the model with diverse, representative data
B) Maintain AI-driven hiring decisions since they optimize efficiency
C) Reduce AI transparency to avoid legal challenges
D) Expand AI hiring automation without addressing bias concerns
B) Conduct a bias audit and retrain the model with diverse, representative data - AI in hiring should ensure fair and unbiased candidate selection.
Which principle does the EU AI Act emphasize for AI decision-making transparency?
A) Explainability and human oversight
B) Fully autonomous AI systems
C) Keeping AI models confidential from users
D) Prioritizing efficiency over fairness
A) Explainability and human oversight - The EU AI Act requires AI models to provide understandable and interpretable decisions.
A multinational retailer integrates an AI-powered facial recognition system for security monitoring in stores. The provider claims that the system complies with EU AI regulations, but the deployer notices that the AI falsely identifies customers as potential shoplifters at a higher rate for certain ethnic groups. What governance step must the deployer take?
A) Halt AI deployment, notify regulatory authorities, and request corrective action from the provider
B) Continue AI operation as long as store security is improved
C) Reduce transparency in AI decision-making to prevent customer complaints
D) Expand AI surveillance without modifying the algorithm
C) Halt AI deployment, notify regulatory authorities, and request corrective action from the provider - Deployers must prevent AI-induced discrimination and ensure compliance.
Which regulation governs AI data privacy and user rights in California?
A) GDPR
B) CCPA
C) HIPAA
D) AI Transparency Act
B) CCPA - The California Consumer Privacy Act (CCPA) regulates AI data privacy and consumer rights.
Which of the following is a key component of AI transparency?
A) Providing users with explanations for AI-generated decisions
B) Keeping AI decision-making confidential to protect intellectual property
C) Removing human intervention from AI systems
D) Increasing algorithmic complexity to improve efficiency
A) Providing users with explanations for AI-generated decisions - Transparency ensures AI decisions can be understood and scrutinized.
What is the best governance approach for mitigating AI bias?
A) Conducting regular audits and retraining AI models with diverse data
B) Reducing AI transparency to avoid scrutiny
C) Ensuring AI models operate without regulatory oversight
D) Maintaining AI models without periodic fairness evaluations
A) Conducting regular audits and retraining AI models with diverse data - Bias mitigation requires ongoing monitoring and adjustment.
Which of the following is a common risk associated with deep learning models?
A) They often lack transparency and explainability
B) They are always free of bias
C) They require no oversight once deployed
D) They cannot be used for decision-making
A) They often lack transparency and explainability - Deep learning models can be complex and difficult to interpret, posing accountability challenges.
A large corporation uses AI-driven recruitment software to screen job applicants. After deployment, a third-party audit finds that the AI system disproportionately rejects female candidates for leadership positions, reinforcing historical hiring biases. What governance action should the company take?
A) Conduct a fairness audit and adjust AI decision-making criteria to eliminate bias
B) Maintain AI-driven hiring decisions as they optimize efficiency
C) Reduce transparency in AI hiring decisions to prevent public scrutiny
D) Expand AI automation in recruitment without reviewing fairness concerns
D) Conduct a fairness audit and adjust AI decision-making criteria to eliminate bias - AI hiring tools must be equitable and free from discrimination.
What is a fundamental principle of responsible AI governance?
A) Ensuring AI systems do not cause harm to individuals or society
B) Maximizing AI performance without regulation
C) Allowing AI to operate with complete autonomy
D) Reducing transparency in AI decision-making
A) Ensuring AI systems do not cause harm to individuals or society - Responsible AI prioritizes ethical considerations to prevent negative impacts.
A company deploys an AI model to predict employee productivity based on keystroke patterns and email activity. Employees raise concerns that the system violates privacy rights and creates an unfair surveillance environment. What is the BEST governance response?
A) Continue using the system since it improves productivity tracking
B) Conduct a privacy impact assessment and consider alternative performance evaluation methods
C) Increase AI model complexity to avoid employee detection
D) Restrict employee access to personal communication platforms
B) Conduct a privacy impact assessment and consider alternative performance evaluation methods - AI monitoring in the workplace must comply with privacy and ethical guidelines.
Which of the following is a key responsibility of AI deployers under the EU AI Act?
A) Monitoring AI system performance and reporting incidents to market surveillance authorities
B) Conducting pre-market conformity assessments
C) Granting AI providers approval to deploy AI systems
D) Developing technical AI documentation
D) Monitoring AI system performance and reporting incidents to market surveillance authorities - Deployers must ensure safe and compliant AI usage.
How does the ‘Measure’ function in the NIST AI RMF contribute to AI risk management?
A) It provides metrics to assess AI model performance, bias, and security risks
B) It removes the need for human oversight in AI governance
C) It ensures AI models operate without regulatory audits
D) It prioritizes speed over safety in AI systems
A) It provides metrics to assess AI model performance, bias, and security risks - Measurement is key to tracking AI risks and effectiveness.
What is a primary risk of using black-box AI models?
A) Lack of transparency and interpretability
B) Increased efficiency and reduced bias
C) Elimination of the need for human oversight
D) Exemption from regulatory requirements
D) Lack of transparency and interpretability - Black-box models make it difficult to assess fairness and accountability.
An AI-driven hiring platform is designed to screen and rank job applicants. After a regulatory investigation, it is found that the model disproportionately favors male candidates for leadership roles. What is the MOST responsible course of action?
A) Shut down the model and avoid AI in hiring decisions
B) Conduct bias testing, adjust model training data, and introduce fairness constraints
C) Hide gender-related variables from the dataset and deploy the model without further adjustments
D) Rely solely on human hiring decisions instead of AI screening
B) Conduct bias testing, adjust model training data, and introduce fairness constraints - AI hiring tools must be monitored for bias and fairness to comply with regulations.
What is a primary limitation of relying solely on historical data for training AI models?
A) It ensures fairness across all demographics
B) It increases AI system latency
C) It may reinforce existing biases and societal inequities
D) It prevents data drift over time
C) It may reinforce existing biases and societal inequities - AI trained on biased historical data can perpetuate unfair outcomes.
A university deploys an AI grading system to assess student assignments. However, students notice that certain writing styles are consistently scored lower. What governance step should be taken?
A) Conduct an audit to determine if the AI model exhibits systemic bias
B) Trust the AI system’s decisions since it ensures standardization
C) Remove human grading oversight to prevent interference
D) Reduce transparency in AI grading methods to avoid challenges
A) Conduct an audit to determine if the AI model exhibits systemic bias - AI in education must be fair and free from unintended biases.
A university implements AI-based admissions screening. A study finds that students from rural backgrounds have lower acceptance rates due to AI rankings. What governance measure should be implemented?
A) Conduct an equity audit and adjust AI decision-making criteria
B) Maintain AI-driven admissions as they improve efficiency
C) Reduce AI transparency in admissions criteria
D) Automate all admissions decisions without fairness considerations
A) Conduct an equity audit and adjust AI decision-making criteria - AI in education must promote fair access for all applicants.
A self-driving car’s AI system must decide between hitting a pedestrian or swerving into another vehicle in an unavoidable accident. What governance principle is MOST relevant in designing this decision-making process?
A) Transparency
B) Ethical AI decision-making frameworks
C) Data privacy
D) Cost efficiency
B) Ethical AI decision-making frameworks - Autonomous AI must incorporate ethical considerations for high-stakes decisions.
A school district implements an AI-driven grading system for essays. Teachers notice that creative writing pieces receive lower scores compared to structured responses. What is the MOST responsible governance action?
A) Adjust the AI grading model to account for diverse writing styles
B) Maintain the current grading system to ensure consistency
C) Remove human oversight to prevent grading bias
D) Reduce AI transparency to limit student appeals
A) Adjust the AI grading model to account for diverse writing styles - AI grading must be inclusive and fair for different writing approaches.
Which AI regulation primarily governs data privacy and user rights in the European Union?
A) GDPR
B) CCPA
C) HIPAA
D) AI Fairness Act
A) GDPR - The General Data Protection Regulation (GDPR) sets strict guidelines on AI data privacy, user consent, and accountability.
A financial institution uses AI for credit risk assessments. An independent audit finds that self-employed applicants are disproportionately denied loans. What governance action should the institution take?
A) Reevaluate AI decision criteria to ensure fair lending practices
B) Maintain current AI-driven lending decisions as they minimize financial risk
C) Reduce AI transparency in credit approvals to prevent legal challenges
D) Expand AI automation in lending without modifications
A) Reevaluate AI decision criteria to ensure fair lending practices - AI financial tools should not discriminate based on employment type.
An AI-based social media moderation tool is found to censor posts from marginalized communities more frequently than others. What governance measure should be prioritized?
A) Conduct fairness audits and retrain the model to ensure equal enforcement
B) Maintain the current system since it reduces misinformation overall
C) Reduce transparency in moderation practices to avoid criticism
D) Expand AI enforcement across all content categories without addressing bias
A) Conduct fairness audits and retrain the model to ensure equal enforcement - AI moderation systems must balance fairness and free expression.
Which NIST AI RMF function involves implementing AI risk mitigation strategies?
A) Manage
B) Govern
C) Measure
D) Map
A) Manage - The ‘Manage’ function ensures organizations take proactive steps to address AI risks.
A financial services company implements an AI system for fraud detection. Customers complain that legitimate transactions are being blocked more frequently for specific demographic groups. What should the company do?
A) Increase AI automation to reduce manual review times
B) Conduct a bias audit and adjust the fraud detection algorithm
C) Reduce AI transparency to protect company security policies
D) Implement stricter fraud thresholds without reviewing bias concerns
B) Conduct a bias audit and adjust the fraud detection algorithm - AI fraud detection systems must balance security with fairness.
Which of the following AI applications falls under the category of high-risk AI systems in the EU AI Act?
A) AI systems used for creditworthiness assessment and hiring decisions
B) AI chatbots used for customer support
C) AI-generated entertainment recommendations
D) AI-based grammar correction tools
C) AI systems used for creditworthiness assessment and hiring decisions - These applications impact fundamental rights and require strict compliance measures.
Which of the following is a critical component of a successful AI governance strategy?
A) Aligning AI policies with ethical guidelines and regulatory requirements
B) Ensuring AI models remain proprietary and confidential
C) Eliminating transparency to avoid legal challenges
D) Allowing AI to function autonomously without oversight
A) Aligning AI policies with ethical guidelines and regulatory requirements - AI governance strategies must be ethical and legally compliant.
Which regulation mandates the right to explanation for AI-driven decisions affecting individuals?
A) GDPR
B) CCPA
C) HIPAA
D) AI Ethics Act
A) GDPR - The General Data Protection Regulation ensures individuals have the right to understand AI decisions affecting them.
A self-driving car manufacturer deploys an AI model for lane detection, but researchers discover it is vulnerable to adversarial patches—small, carefully designed modifications to road signs that mislead the AI into making dangerous decisions. How should the manufacturer respond?
A) Implement adversarial robustness techniques and retrain AI with diverse adversarial examples
B) Maintain AI-driven lane detection since it improves efficiency in standard conditions
C) Reduce AI transparency to prevent security vulnerabilities from being exposed
D) Expand self-driving AI deployment without additional robustness measures
D) Implement adversarial robustness techniques and retrain AI with diverse adversarial examples - AI in autonomous vehicles must be resilient against manipulation.
Which of the following is a major security risk in AI systems?
A) Adversarial attacks that manipulate AI model predictions
B) The ability of AI to function without human intervention
C) AI models that require human oversight
D) AI systems that are transparent and explainable
A) Adversarial attacks that manipulate AI model predictions - Security risks in AI include attacks that trick models into making incorrect decisions.
Which of the following is an example of an AI system categorized as ‘high-risk’ under the EU AI Act?
A) AI-powered music streaming suggestions
B) AI in autonomous vehicles
C) AI that generates news articles
D) AI for social media content moderation
B) AI in autonomous vehicles - High-risk AI includes systems that impact safety and fundamental rights, such as autonomous vehicles.
A city government implements an AI-powered predictive policing system. A public investigation reveals that the AI disproportionately flags low-income neighborhoods as high-crime areas, leading to increased policing in those communities. What governance action should be taken?
A) Conduct a fairness audit, retrain AI models, and introduce human oversight
B) Maintain AI-driven policing decisions as they improve crime detection
C) Reduce AI transparency to prevent public concern
D) Expand AI-driven policing without addressing fairness concerns
B) Conduct a fairness audit, retrain AI models, and introduce human oversight - AI policing tools should be equitable and unbiased.
A city government deploys an AI-driven traffic control system designed to optimize congestion. However, after deployment, reports indicate that the system has led to increased pollution in lower-income neighborhoods and disproportionately directs commercial traffic away from wealthier areas. Furthermore, an audit finds that the AI model does not factor in environmental sustainability. How should the city respond?
A) Adjust AI optimization criteria to balance traffic flow equitably and consider environmental impact
B) Maintain AI-driven traffic routing since it improves efficiency for most drivers
C) Reduce AI transparency to prevent opposition from local communities
D) Expand AI-driven traffic control without reviewing geographic fairness concerns
B) Adjust AI optimization criteria to balance traffic flow equitably and consider environmental impact - AI infrastructure must ensure fairness and sustainability.
What distinguishes machine learning from traditional rule-based programming?
A) Machine learning enables systems to learn from data and improve over time without explicit programming
B) Machine learning operates solely based on pre-written rules and conditions
C) Machine learning does not require data to make decisions
D) Machine learning models cannot be retrained after deployment
B) Machine learning enables systems to learn from data and improve over time without explicit programming - Unlike rule-based systems, ML adapts dynamically based on experience.
A predictive AI system for school dropout risk incorrectly labels many minority students as at-risk. What governance measure should be taken?
A) Conduct an impact assessment and adjust the model for fairness
B) Maintain current AI predictions since they improve school intervention rates
C) Reduce AI transparency to prevent parental concerns
D) Automate all school intervention decisions to remove human bias
A) Conduct an impact assessment and adjust the model for fairness - AI in education must not reinforce biases against certain groups.
What is the main risk of model drift in AI governance?
A) AI decision accuracy may degrade over time
B) AI models become more transparent and explainable
C) AI efficiency improves without oversight
D) AI systems require less training data over time
A) AI decision accuracy may degrade over time - Model drift occurs when AI systems deviate from expected outcomes due to changing data patterns.
What is the main goal of AI model explainability?
A) To ensure AI models remain confidential and proprietary
B) To provide understandable reasons for AI-driven decisions
C) To remove the need for regulatory oversight
D) To increase the speed of AI decision-making
C) To provide understandable reasons for AI-driven decisions - Explainability is key to responsible AI governance.
A city uses AI-powered traffic cameras to detect violations. Residents report that the AI disproportionately issues tickets in certain neighborhoods. What governance step should be taken?
A) Review the dataset for biases and adjust the AI system accordingly
B) Maintain current enforcement since overall violations have decreased
C) Reduce public access to traffic enforcement data to prevent scrutiny
D) Allow AI to issue fines without human oversight
A) Review the dataset for biases and adjust the AI system accordingly - AI enforcement must be fair and not disproportionately target specific communities.
Which of the following is the PRIMARY goal of AI governance frameworks?
A) Maximizing AI system profitability
B) Ensuring AI aligns with ethical, legal, and social norms
C) Replacing human decision-making entirely
D) Preventing AI system innovation
B) Ensuring AI aligns with ethical, legal, and social norms - AI governance frameworks help balance innovation with risk mitigation.
A hospital AI system predicts patient deterioration risks but is found to over-prioritize ICU admissions for certain demographics. What is the most ethical governance step?
A) Conduct a bias audit and recalibrate the AI model
B) Maintain the current AI system since it maximizes patient survival rates
C) Reduce AI transparency in ICU recommendations to prevent legal issues
D) Allow AI to fully automate ICU admissions without human review
A) Conduct a bias audit and recalibrate the AI model - AI in healthcare should ensure equitable treatment for all patients.
A university deploys an AI-driven proctoring system to monitor online exams. Students raise concerns that it falsely flags neurodiverse students for suspicious behavior. What governance measure should be prioritized?
A) Review and modify AI behavior detection criteria to accommodate diverse needs
B) Maintain the current model as it improves exam integrity
C) Reduce AI transparency to prevent appeals
D) Expand automated monitoring without reviewing accuracy
A) Review and modify AI behavior detection criteria to accommodate diverse needs - AI-based surveillance in education should be inclusive and fair.
Which AI application is MOST likely to require a Data Protection Impact Assessment (DPIA) under GDPR?
A) AI-powered music recommendation engine
B) AI used in automated hiring decisions
C) AI that generates stock market predictions
D) AI that translates languages
B) AI used in automated hiring decisions - AI that affects individual rights and processes sensitive data requires a DPIA under GDPR.
An AI-powered résumé screening tool is found to reject significantly more applications from candidates with non-Western names. What is the MOST effective governance response?
A) Remove all name-related information from the training data and redeploy the model
B) Conduct a bias audit, retrain the model with diverse data, and implement fairness constraints
C) Maintain the model’s current training data to avoid altering hiring patterns
D) Shift hiring decisions entirely to AI for more objective results
B) Conduct a bias audit, retrain the model with diverse data, and implement fairness constraints - Fair AI hiring requires bias mitigation and compliance with equal opportunity laws.
A city government deploys an AI-driven traffic control system. A study finds that congestion is worsening in lower-income neighborhoods. What governance step should be prioritized?
A) Adjust AI traffic algorithms to distribute traffic flow more equitably
B) Maintain current AI traffic control since it improves efficiency overall
C) Reduce transparency in AI decision-making to prevent opposition
D) Expand AI-driven traffic control without reviewing bias factors
D) Adjust AI traffic algorithms to distribute traffic flow more equitably - AI in public infrastructure should benefit all communities equally.
Which regulation mandates the right to explanation for AI-driven decisions affecting individuals?
A) GDPR
B) CCPA
C) HIPAA
D) The EU Cybersecurity Act
D) GDPR - The General Data Protection Regulation ensures individuals have the right to understand AI decisions affecting them.
Under the EU AI Act, what must deployers of AI systems do if they detect risks or malfunctions post-deployment?
A) Report incidents to market surveillance authorities and take corrective action
B) Continue using the AI system while waiting for provider guidance
C) Modify the AI system’s functionality without notifying regulators
D) Cease AI system operation without providing further documentation
C) Report incidents to market surveillance authorities and take corrective action - Deployers are responsible for identifying and mitigating risks in deployed AI.
Which core function of the NIST AI RMF ensures ongoing oversight and compliance in AI risk management?
A) Govern
B) Map
C) Measure
D) Manage
A) Govern - The ‘Govern’ function ensures that AI risk management remains an ongoing priority in organizations.
A financial institution deploys an AI-driven credit approval system. However, an independent review finds that the AI disproportionately rejects loan applications from small business owners in low-income areas due to biases in historical lending data. What governance measure should be implemented?
A) Conduct a fairness audit and recalibrate AI decision criteria
B) Maintain AI-driven lending decisions as they optimize financial risk
C) Reduce AI transparency in loan approvals to prevent legal challenges
D) Expand AI automation in lending without modifications
D) Conduct a fairness audit and recalibrate AI decision criteria - AI in finance should ensure fair lending practices.
An AI-based hiring tool used in the EU is found to systematically rank male candidates higher than female candidates for leadership roles. What governance measure should be prioritized?
A) Conduct a bias audit and recalibrate AI hiring criteria
B) Maintain current AI rankings as they align with historical hiring trends
C) Reduce AI transparency in hiring decisions to prevent scrutiny
D) Automate all hiring decisions without reviewing fairness concerns
A) Conduct a bias audit and recalibrate AI hiring criteria - AI recruitment systems must comply with EU fairness and anti-discrimination laws.
A predictive policing AI system is deployed in a city and leads to higher arrest rates in historically marginalized communities. What governance measure should be implemented?
A) Conduct an independent audit to assess bias and adjust AI decision-making
B) Maintain current AI enforcement practices since they increase efficiency
C) Expand AI surveillance to all communities equally without reviewing existing bias
D) Reduce public access to AI crime predictions to prevent criticism
A) Conduct an independent audit to assess bias and adjust AI decision-making - AI policing tools must be fair and avoid reinforcing historical biases.
What is the main purpose of AI impact assessments?
A) Evaluating and mitigating potential ethical, legal, and social risks of AI deployment
B) Removing regulatory constraints to accelerate AI development
C) Increasing AI decision-making complexity
D) Eliminating the need for human intervention in AI systems
B) Evaluating and mitigating potential ethical, legal, and social risks of AI deployment - AI impact assessments help organizations proactively address risks.
An AI healthcare system is designed to predict disease risk based on patient data. However, patients express concerns about privacy violations. What is the BEST way to address these concerns while maintaining AI effectiveness?
A) Implement differential privacy techniques and allow opt-out options
B) Continue using the AI model since medical accuracy outweighs privacy concerns
C) Disable explainability features to protect patient confidentiality
D) Limit AI-based diagnostics to only high-risk patients
A) Implement differential privacy techniques and allow opt-out options - Privacy-enhancing techniques help protect patient data while ensuring AI effectiveness.
An e-commerce platform deploys an AI fraud detection system that frequently flags transactions from certain geographic regions. This results in legitimate transactions being blocked. What should the company do?
A) Adjust the AI model by incorporating diverse transaction patterns
B) Increase penalties for flagged users to deter fraud
C) Remove human oversight to make fraud detection fully automated
D) Block all transactions from flagged regions permanently
A) Adjust the AI model by incorporating diverse transaction patterns - AI fraud detection must be balanced to prevent discriminatory outcomes.
A healthcare AI system prioritizes treatment recommendations based on patient age, unintentionally favoring younger patients over older ones. What governance step should be prioritized?
A) Review and adjust AI decision criteria to ensure ethical fairness
B) Maintain AI-driven recommendations as they improve efficiency
C) Reduce transparency in AI medical decision-making to avoid scrutiny
D) Automate all medical diagnoses without human intervention
A) Review and adjust AI decision criteria to ensure ethical fairness - AI in healthcare must be equitable for all patients.
What is the primary risk of using biased training data in an AI system?
A) Reduced computational efficiency
B) Increased model accuracy
C) Discriminatory or unfair outcomes
D) Faster processing speeds
C) Discriminatory or unfair outcomes - Biased training data can lead to AI systems producing unfair and discriminatory results.
What is the function of machine learning in AI systems?
A) To allow AI models to improve their performance by learning from data
B) To eliminate the need for human decision-making
C) To operate entirely without data input
D) To ensure all AI decisions are explainable
A) To allow AI models to improve their performance by learning from data - Machine learning enables AI models to recognize patterns and make predictions.
What does the EU AI Act require for biometric identification systems?
A) Strict regulatory approval before deployment
B) Full automation without human intervention
C) No restrictions, as biometric AI is considered low-risk
D) Elimination of transparency in biometric decision-making
A) Strict regulatory approval before deployment - The EU AI Act imposes strong compliance measures on biometric AI to prevent misuse.
Which principle of AI governance ensures that AI systems provide explanations for their decisions?
A) Transparency
B) Model efficiency
C) Data minimization
D) Algorithmic secrecy
A) Transparency - Explainability in AI decision-making is key to trust and accountability.
Under the EU AI Act, which category of AI poses the GREATEST risk and is banned?
A) High-risk AI
B) Limited-risk AI
C) Unacceptable-risk AI
D) Minimal-risk AI
C) Unacceptable-risk AI - The EU AI Act prohibits AI systems such as social credit scoring and biometric surveillance due to risks to fundamental rights.
What is model drift in AI systems?
A) When AI model performance degrades over time due to changing data
B) When AI models become more transparent with usage
C) When AI systems require no further updates after deployment
D) When AI decision-making becomes more efficient
D) When AI model performance degrades over time due to changing data - AI governance requires monitoring to prevent unintended consequences.
An AI-powered predictive policing system in the EU is found to disproportionately target specific ethnic groups. What is the most responsible governance action under the EU AI Act?
A) Conduct a fairness audit and adjust AI policing decision criteria
B) Maintain AI-driven crime prediction as it enhances law enforcement
C) Reduce transparency in AI policing models to prevent public opposition
D) Expand AI-driven policing decisions without reviewing fairness concerns
A) Conduct a fairness audit and adjust AI policing decision criteria - The EU AI Act mandates fairness in AI decision-making, especially in law enforcement.
What is the primary role of an AI ethics committee in governance frameworks?
A) To oversee AI deployments and ensure compliance with ethical standards
B) To optimize AI systems for maximum efficiency
C) To reduce regulatory oversight for AI systems
D) To ensure AI models operate without human intervention
A) To oversee AI deployments and ensure compliance with ethical standards - AI ethics committees help guide responsible AI development.
An AI-driven content recommendation system on a video-sharing platform starts amplifying conspiracy theories due to high engagement levels. What governance strategy should the platform implement?
A) Introduce content verification and reduce algorithmic bias towards misinformation
B) Allow the AI to continue optimizing for engagement without interference
C) Remove transparency in content ranking to prevent public backlash
D) Expand the AI’s recommendations based on existing user behavior
A) Introduce content verification and reduce algorithmic bias towards misinformation - AI content recommendations should prioritize accuracy and responsibility.
A university is considering using AI for grading essays. Faculty members worry that the AI may unfairly penalize creative writing styles. How should the university approach this issue?
A) Implement a human-in-the-loop system for final grading decisions
B) Allow the AI system to grade essays without oversight to maintain consistency
C) Reduce transparency in AI grading methods to prevent appeals
D) Train the AI to favor more formulaic writing styles
A) Implement a human-in-the-loop system for final grading decisions - AI grading should involve human oversight to ensure fairness and adaptability.
Which of the following would be considered a ‘black box’ AI model?
A) A model whose decisions can be easily traced and explained
B) A simple linear regression model with transparent weights
C) A deep neural network with complex decision-making that lacks interpretability
D) A decision tree model with clear branching logic
C) A deep neural network with complex decision-making that lacks interpretability - ‘Black box’ AI refers to models where decision processes are opaque and difficult to explain.
A city government deploys an AI-powered facial recognition system for public security. A civil rights organization reports that the AI model has a higher error rate for identifying individuals from certain racial groups. What is the most responsible governance action?
A) Conduct a fairness audit and retrain the AI model with diverse datasets
B) Maintain AI-driven facial recognition since it enhances public security
C) Reduce AI transparency to avoid public criticism
D) Expand AI facial recognition deployment without reviewing bias concerns
C) Conduct a fairness audit and retrain the AI model with diverse datasets - AI security tools must be fair and unbiased.
Under the EU AI Act, which category of AI systems is considered high-risk?
A) AI applications affecting fundamental rights, health, and safety
B) AI models used in entertainment applications
C) AI-driven chatbots for customer service
D) AI tools used for personal productivity
C) AI applications affecting fundamental rights, health, and safety - High-risk AI systems require strict compliance and oversight.
A financial institution uses an AI model to detect money laundering, but it falsely flags transactions from small businesses more often than large corporations. What governance action should be taken?
A) Conduct a bias assessment and refine the model’s risk criteria
B) Maintain the AI system as is since it detects financial crime effectively
C) Reduce transparency in fraud detection to protect proprietary algorithms
D) Expand AI surveillance without adjusting bias concerns
A) Conduct a bias assessment and refine the model’s risk criteria - AI fraud detection should be unbiased and avoid disproportionately targeting specific groups.
A self-driving vehicle AI system has a higher accident rate in low-light conditions. What governance step should be taken?
A) Improve training data to include more nighttime driving scenarios
B) Maintain current AI behavior since it performs well in daylight
C) Reduce transparency about AI driving performance to avoid liability
D) Automate all driving decisions without human intervention
A) Improve training data to include more nighttime driving scenarios - AI in transportation must be tested and trained in diverse conditions.
Which of the following is a key ethical concern in AI-driven hiring systems?
A) AI models may reinforce biases present in historical hiring data
B) AI hiring eliminates all discrimination
C) AI-driven hiring systems never require human oversight
D) AI systems are immune to legal challenges in employment decisions
A) AI models may reinforce biases present in historical hiring data - AI recruitment should be continuously monitored to prevent discrimination.
Which of the following scenarios presents the HIGHEST risk under AI governance frameworks?
A) An AI chatbot that provides general customer service responses
B) A self-learning AI system used for financial loan approvals
C) An AI-powered search engine ranking algorithm
D) An AI system that generates automated email responses
B) A self-learning AI system used for financial loan approvals - AI in financial services requires strict governance due to potential discrimination and regulatory impact.
What is the role of encryption in AI security?
A) Protecting sensitive AI model data from unauthorized access
B) Reducing AI decision-making speed to improve accuracy
C) Ensuring AI models operate without human oversight
D) Allowing AI systems to run on minimal data
A) Protecting sensitive AI model data from unauthorized access - Encryption helps safeguard AI data from security breaches.
Which of the following best defines AI robustness?
A) The ability of AI systems to perform reliably under different conditions
B) The speed at which an AI model processes data
C) The complexity of an AI algorithm
D) The need for AI to operate without any human involvement
A) The ability of AI systems to perform reliably under different conditions - AI robustness ensures consistent performance across diverse scenarios.
Which term describes when an AI model’s predictions degrade over time due to changing data patterns?
A) Algorithmic fairness
B) Model drift
C) AI optimization
D) Latency reduction
B) Model drift - AI models need retraining when data patterns change significantly over time, affecting accuracy.
A bank implements an AI model to detect fraudulent transactions. Customers from specific ethnic backgrounds report frequent false positives. What is the most responsible governance action?
A) Conduct a bias analysis and adjust the model to minimize discriminatory outcomes
B) Maintain the fraud detection system as is since it reduces overall financial risk
C) Reduce transparency in fraud detection algorithms to prevent regulatory scrutiny
D) Expand fraud detection to all customers equally without bias adjustments
A) Conduct a bias analysis and adjust the model to minimize discriminatory outcomes - AI fraud detection should be fair and not disproportionately impact specific groups.
Which principle of AI governance ensures that AI systems operate within legal and ethical boundaries?
A) Transparency
B) Fairness
C) Accountability
D) Autonomy
C) Accountability - AI accountability ensures organizations take responsibility for AI decisions and their consequences.
What is a key ethical concern with AI-powered predictive policing?
A) It may reinforce systemic biases in law enforcement
B) It completely eliminates crime in all communities
C) AI predictive policing requires no human oversight
D) AI-driven policing decisions are always more accurate than human decisions
D) It may reinforce systemic biases in law enforcement - AI policing should not disproportionately target specific groups.
A financial institution develops an AI-driven chatbot for customer support. However, users report that the chatbot struggles with complex queries, often providing incorrect or misleading answers. The chatbot operates using a natural language processing (NLP) model trained on limited historical customer interactions. What is the best governance action?
A) Expand training data and introduce a hybrid model incorporating human oversight
B) Maintain the AI chatbot as it improves efficiency for simple inquiries
C) Reduce AI transparency to prevent user complaints
D) Automate all customer interactions without human intervention
A) Expand training data and introduce a hybrid model incorporating human oversight - AI chatbots should be trained on diverse queries and include human intervention for complex cases.
What is the role of third-party audits in AI governance?
A) To independently verify compliance with ethical and regulatory standards
B) To replace internal AI governance teams
C) To increase AI system speed
D) To reduce the need for regulatory oversight
A) To independently verify compliance with ethical and regulatory standards - Third-party audits help ensure transparency and accountability.
A social media AI recommendation system prioritizes divisive political content. Regulators warn that this may contribute to misinformation. What governance step should be taken?
A) Adjust the algorithm to prioritize credibility over engagement
B) Maintain current AI content ranking to maximize platform revenue
C) Reduce AI transparency in content moderation to prevent scrutiny
D) Expand AI-driven recommendations without modifications
A) Adjust the algorithm to prioritize credibility over engagement - AI-driven content distribution should promote factual information over sensationalism.
Which AI governance principle ensures that AI systems provide users with understandable reasons for decisions?
A) Efficiency
B) Model drift
C) Data obfuscation
D) Explainability
D) Explainability - AI decisions should be interpretable and transparent to users and stakeholders.
Which of the following is an effective method to increase the robustness of an AI system against adversarial attacks?
A) Training the AI model with adversarial examples to improve resilience
B) Reducing AI transparency to prevent attackers from understanding its weaknesses
C) Increasing AI autonomy without security constraints
D) Preventing all updates to the model after deployment
A) Training the AI model with adversarial examples to improve resilience - This technique strengthens AI defenses against adversarial manipulation.
A government agency uses AI to predict which citizens are eligible for social benefits. A study finds that applicants from lower-income backgrounds face disproportionately high rejection rates. What is the best governance action?
A) Maintain the AI-driven eligibility model as it reduces fraud
B) Conduct a fairness audit and recalibrate AI decision-making
C) Reduce transparency in eligibility criteria to avoid scrutiny
D) Expand AI automation in eligibility assessment without modifications
D) Conduct a fairness audit and recalibrate AI decision-making - AI in public services should be equitable for all users.
Which factor is the MOST important when assessing AI fairness?
A) Dataset diversity
B) AI processing speed
C) Algorithm complexity
D) API response time
A) Dataset diversity - A diverse dataset helps reduce bias and ensures fairness in AI decision-making.
A self-driving car manufacturer finds that its AI system is less accurate in detecting pedestrians at night. What governance step should be prioritized?
A) Improve training data and retrain the model to handle low-light scenarios
B) Maintain AI behavior since it performs well in daytime conditions
C) Reduce transparency about AI weaknesses to prevent liability concerns
D) Automate all driving decisions without further testing
D) Improve training data and retrain the model to handle low-light scenarios - AI safety requires continuous training in diverse conditions.
What is the role of bias audits in AI governance?
A) To increase system performance speed
B) To detect and mitigate unfair bias in AI models
C) To reduce the need for human oversight
D) To decrease AI training time
B) To detect and mitigate unfair bias in AI models - Bias audits help ensure fairness and prevent discrimination in AI applications.
A smart city deploys AI-powered cameras for real-time traffic enforcement. Reports indicate that the system disproportionately fines vehicles from certain neighborhoods. What governance step should be prioritized?
A) Audit the AI model and retrain it with a more representative dataset
B) Maintain the current system since overall traffic violations have decreased
C) Reduce transparency to prevent public scrutiny
D) Expand the AI surveillance system without reviewing biases
A) Audit the AI model and retrain it with a more representative dataset - AI enforcement should be fair and free from geographic biases.
Why is continuous monitoring of AI models necessary after deployment?
A) To increase system processing speeds
B) To detect and mitigate potential bias, drift, or errors
C) To reduce costs associated with AI compliance
D) To allow AI models to operate without human intervention
B) To detect and mitigate potential bias, drift, or errors - Continuous monitoring ensures AI systems remain fair, accurate, and compliant over time.
Which principle of AI governance ensures users can challenge AI-driven decisions?
A) Contestability
B) Automation
C) Complexity
D) Secrecy
B) Contestability - AI governance should allow users to question and appeal AI-generated outcomes.
What is the MAIN reason why explainability is important in AI governance?
A) It improves computational efficiency
B) It allows stakeholders to understand how AI makes decisions
C) It eliminates the need for risk assessments
D) It reduces AI system costs
B) It allows stakeholders to understand how AI makes decisions - Explainability increases trust, accountability, and compliance in AI systems.
A healthcare AI model is found to disproportionately underdiagnose heart disease in female patients. What governance action should be prioritized?
A) Conduct a bias audit and retrain the model with more diverse data
B) Maintain the AI model since it performs well for male patients
C) Reduce transparency in AI decision-making to prevent scrutiny
D) Automate all medical diagnoses without human intervention
A) Conduct a bias audit and retrain the model with more diverse data - AI in healthcare must be trained on datasets representative of all demographics.
A government deploys AI-based social welfare distribution. Analysts find that the AI system assigns lower benefit scores to applicants with gaps in employment history, disproportionately affecting caregivers who took time off for family responsibilities. What governance response should be prioritized?
A) Conduct an equity audit and recalibrate AI decision criteria
B) Maintain AI-driven welfare assessments as they improve efficiency
C) Reduce AI transparency to prevent policy challenges
D) Expand AI automation in social services without reviewing fairness concerns
B) Conduct an equity audit and recalibrate AI decision criteria - AI in public service distribution should prevent unfair penalization.
What is model explainability in AI governance?
A) The ability for stakeholders to understand and interpret AI decision-making
B) The process of making AI models more efficient at the cost of transparency
C) The elimination of the need for regulatory oversight
D) The use of AI models that operate without any human intervention
A) The ability for stakeholders to understand and interpret AI decision-making - Explainability ensures transparency and accountability in AI systems.
What is the primary purpose of an AI ethics committee in an organization?
A) To oversee AI deployment and ensure compliance with ethical standards
B) To maximize AI efficiency without regulation
C) To ensure AI models remain confidential and proprietary
D) To remove human oversight from AI decision-making
B) To oversee AI deployment and ensure compliance with ethical standards - Ethics committees play a key role in AI governance and risk management.
A hospital AI system prioritizes younger patients for ICU admissions. Critics argue this disadvantages elderly patients. What is the most ethical governance response?
A) Maintain current AI triage as it optimizes hospital resources
B) Review and adjust the AI model to ensure equitable treatment
C) Reduce AI transparency to prevent ethical concerns
D) Expand AI-driven patient prioritization without modifications
B) Review and adjust the AI model to ensure equitable treatment - AI healthcare systems should not unfairly disadvantage certain patient groups.
An AI-powered home loan approval system frequently denies loans in historically redlined areas. Regulators suspect algorithmic discrimination. What is the MOST ethical governance response?
A) Conduct a bias audit and implement fairness constraints in lending decisions
B) Maintain the current system since it aligns with historical lending patterns
C) Reduce transparency in lending decisions to prevent external scrutiny
D) Expand AI-based credit assessments without adjusting discrimination factors
A) Conduct a bias audit and implement fairness constraints in lending decisions - AI in finance should not perpetuate historical discrimination.
What is the key risk associated with AI-driven facial recognition in law enforcement?
A) Increased model efficiency
B) Risk of false positives and biased identification
C) Reduced computational costs
D) Increased public trust
B) Risk of false positives and biased identification - AI facial recognition must be rigorously tested to prevent wrongful identification.
Which of the following is a key factor in responsible AI deployment?
A) Regular auditing and impact assessments
B) Removing regulatory barriers to AI growth
C) Ensuring AI operates without human oversight
D) Keeping AI decision-making confidential
A) Regular auditing and impact assessments - Ongoing audits help organizations ensure AI fairness, safety, and compliance.
An AI-driven financial risk assessment model is found to deny loans to female entrepreneurs at a higher rate than male applicants. What governance step should be prioritized?
A) Conduct a bias analysis and adjust AI decision criteria for fairness
B) Maintain AI-driven lending decisions to optimize financial risk assessments
C) Reduce AI transparency to prevent public concerns
D) Automate all loan approval decisions without human review
A) Conduct a bias analysis and adjust AI decision criteria for fairness - AI financial services should not reinforce gender-based disparities.
A healthcare provider deploys an AI diagnostic tool that outperforms human doctors in some cases but lacks explainability. Which ethical principle is MOST at risk?
A) Autonomy
B) Privacy
C) Transparency
D) Beneficence
C) Transparency - Lack of explainability in high-stakes AI applications can reduce trust and accountability.
A self-driving car manufacturer receives complaints that its AI system struggles to recognize pedestrians in wheelchairs. What governance step should be taken?
A) Maintain AI behavior as it works well in standard scenarios
B) Increase training data diversity and conduct fairness audits
C) Reduce transparency about AI weaknesses to avoid liability
D) Automate all driving decisions without further human intervention
C) Increase training data diversity and conduct fairness audits - AI in transportation must be designed to recognize all users safely.
A multinational corporation uses AI to monitor employee productivity. Employees raise privacy concerns about constant surveillance. What governance measure should the company implement?
A) Introduce clear policies on AI monitoring and provide opt-out options
B) Continue monitoring without disclosure to maintain efficiency
C) Reduce transparency about AI tracking to avoid legal scrutiny
D) Automate employee monitoring entirely to remove human oversight
A) Introduce clear policies on AI monitoring and provide opt-out options - AI workplace monitoring should balance efficiency with privacy rights.
A municipality deploys an AI-powered traffic monitoring system to optimize traffic flow. After six months, residents notice that the system prioritizes certain neighborhoods, leading to increased congestion in lower-income areas. How should city officials respond?
A) Maintain the system as is since traffic efficiency is maximized overall
B) Conduct an equity impact assessment and retrain the AI with more balanced data
C) Disable the AI system and revert to manual traffic management
D) Increase law enforcement presence in affected areas to mitigate congestion
B) Conduct an equity impact assessment and retrain the AI with more balanced data - AI in public infrastructure must consider fairness and avoid unintended negative impacts.
A hiring AI tool consistently ranks candidates from certain universities higher, leading to reduced diversity in hiring decisions. What governance step should the company take?
A) Conduct a bias audit and recalibrate AI decision-making criteria
B) Maintain AI hiring practices since they reflect historical success
C) Reduce transparency in AI hiring to prevent legal challenges
D) Expand AI automation in hiring without reviewing fairness concerns
A) Conduct a bias audit and recalibrate AI decision-making criteria - AI hiring tools should ensure fair evaluation across all candidates.
A financial AI system denies loans to customers with non-traditional employment histories. What should the company do?
A) Review training data and adjust AI decision criteria for fairness
B) Maintain the model since it reflects historical lending risk assessments
C) Reduce transparency in AI loan approvals to prevent challenges
D) Automate all lending decisions without reviewing AI outcomes
A) Review training data and adjust AI decision criteria for fairness - AI lending should not disadvantage individuals based on employment type.
What is the primary concern of fairness in AI governance?
A) Preventing AI systems from producing discriminatory or biased outcomes
B) Ensuring AI operates as quickly as possible
C) Reducing human oversight to increase efficiency
D) Allowing AI models to function without audits
A) Preventing AI systems from producing discriminatory or biased outcomes - AI fairness ensures equitable treatment across all demographic groups.
An AI-powered medical diagnostics tool is found to have a significantly lower accuracy rate for rare diseases due to data sparsity. What is the most responsible governance action?
A) Expand training datasets and improve AI learning models
B) Maintain AI-driven diagnostics as they improve overall efficiency
C) Reduce AI transparency to avoid public concern
D) Automate all medical diagnoses without human review
C) Expand training datasets and improve AI learning models - AI in healthcare should be inclusive and effective for all conditions.
Which of the following best describes algorithmic fairness in AI governance?
A) The principle that AI should not produce discriminatory outcomes
B) AI models should always prioritize efficiency over fairness
C) AI systems should make decisions without transparency
D) AI should not be subject to regulatory audits
D) The principle that AI should not produce discriminatory outcomes - Algorithmic fairness ensures that AI does not reinforce bias in decision-making.
Which governance principle ensures that AI models can be inspected and understood by regulators and users?
A) Explainability
B) Data obfuscation
C) Algorithmic opacity
D) Computational efficiency
A) Explainability - AI systems should be interpretable to ensure accountability and compliance.
What is a primary limitation of relying solely on historical data for training AI models?
A) It ensures fairness across all demographics
B) It increases AI system latency
C) It may reinforce existing biases and societal inequities
D) It prevents data drift over time
C) It may reinforce existing biases and societal inequities - AI trained on biased historical data can perpetuate unfair outcomes.
A healthcare AI system is found to misdiagnose certain racial groups at a higher rate than others. What is the most responsible governance action?
A) Conduct a bias analysis and retrain the model using diverse datasets
B) Maintain the current system since it performs well for the majority of cases
C) Reduce transparency to prevent public concern
D) Automate all diagnoses without human review
A) Conduct a bias analysis and retrain the model using diverse datasets - AI in healthcare must be fair and equitable for all patient demographics.
Which of the following is a key requirement for AI explainability in high-risk AI applications under the EU AI Act?
A) AI decisions must be interpretable and justifiable to affected users
B) AI models must be entirely open-source to ensure full transparency
C) AI systems must always operate without human involvement
D) AI providers are required to keep decision-making processes confidential
A) AI decisions must be interpretable and justifiable to affected users - Explainability ensures AI-driven decisions can be understood and challenged.
A financial AI model is found to deny loans to self-employed applicants at a disproportionately high rate. What is the best governance response?
A) Review AI decision criteria and ensure fair treatment for all applicants
B) Maintain AI-driven lending decisions as they optimize financial risk
C) Reduce AI transparency in loan approvals to prevent regulatory intervention
D) Expand AI automation in lending without adjustments
A) Review AI decision criteria and ensure fair treatment for all applicants - AI financial systems should not discriminate based on employment type.
A university uses AI to analyze student learning patterns. A review finds that the AI system disproportionately flags non-native English speakers as struggling students. What is the best governance response?
A) Conduct a bias audit and refine AI learning assessment models
B) Maintain AI-driven student evaluations as they improve efficiency
C) Reduce transparency in AI assessments to prevent student complaints
D) Expand AI-driven learning assessments without reviewing fairness concerns
A) Conduct a bias audit and refine AI learning assessment models - AI in education should ensure fairness and inclusivity for all students.
Which regulatory authority is primarily responsible for overseeing compliance with the EU AI Act across member states?
A) The national competent authority in each EU country
B) The AI deployer’s internal compliance team
C) The European Parliament directly
D) The AI system manufacturer
B) The national competent authority in each EU country - Each country enforces AI Act compliance through its designated authority.
Which of the following is a major compliance requirement for AI in high-risk applications?
A) Conducting regular audits, documentation, and human oversight
B) Allowing AI to operate autonomously without monitoring
C) Eliminating transparency requirements to protect proprietary models
D) Reducing regulatory obligations to encourage rapid deployment
D) Conducting regular audits, documentation, and human oversight - Compliance ensures responsible AI use in critical areas.
A self-driving car manufacturer discovers its AI model struggles with detecting pedestrians in low-light conditions. What governance step should be taken?
A) Improve training data and retrain the AI model for night-time driving scenarios
B) Maintain current AI behavior since it performs well in daylight
C) Reduce transparency about AI weaknesses to prevent liability concerns
D) Allow AI to continue driving without further testing
D) Improve training data and retrain the AI model for night-time driving scenarios - AI in transportation must be safe in all conditions.
An AI-powered translation system performs worse for languages with fewer training data examples. What governance step should be taken?
A) Improve dataset diversity and retrain the model for better representation
B) Maintain the current model since it works well for widely spoken languages
C) Reduce transparency in AI performance results to avoid criticism
D) Allow AI to continue evolving without intervention
A) Improve dataset diversity and retrain the model for better representation - AI language models must be trained on diverse datasets to ensure accuracy.
An AI-powered hiring tool is trained on historical company data. Over time, it rejects a disproportionate number of female candidates. What is the BEST course of action?
A) Conduct a bias audit and introduce fairness constraints in the AI model
B) Keep the system unchanged since it reflects past hiring success
C) Reduce transparency in AI hiring decisions to prevent challenges
D) Remove gender-related variables without analyzing their indirect effects
A) Conduct a bias audit and introduce fairness constraints in the AI model - AI hiring must be actively monitored to prevent discrimination.
A university deploys an AI system to detect plagiarism. However, it frequently flags non-native English speakers’ work as plagiarized. What governance measure should be prioritized?
A) Improve the AI training dataset to account for diverse writing styles
B) Remove human oversight to ensure consistency
C) Allow AI to continue flagging submissions without intervention
D) Reduce transparency in plagiarism detection to prevent appeals
A) Improve the AI training dataset to account for diverse writing styles - AI tools in education must be designed to ensure fairness across all students.
Which international organization has established guidelines for AI risk management?
A) World Trade Organization (WTO)
B) International Monetary Fund (IMF)
C) National Institute of Standards and Technology (NIST)
D) Intergovernmental Panel on Climate Change (IPCC)
C) National Institute of Standards and Technology (NIST) - NIST has developed the AI Risk Management Framework to guide responsible AI development.
A government agency deploys an AI surveillance system that was assessed as high-risk under the EU AI Act. After deployment, civil rights groups raise concerns about potential privacy violations. What governance step should be prioritized?
A) Conduct a post-market compliance review and adjust the AI model to align with regulatory requirements
B) Maintain AI-driven surveillance since it enhances security
C) Reduce transparency in AI surveillance decision-making to prevent legal challenges
D) Expand AI deployment without addressing ethical concerns
D) Conduct a post-market compliance review and adjust the AI model to align with regulatory requirements - AI governance must ensure compliance with privacy and human rights protections.
A ride-hailing platform’s AI pricing model increases fares during extreme weather events, disproportionately impacting low-income users. What governance step should be prioritized?
A) Implement ethical pricing caps to prevent exploitative surge pricing
B) Maintain the AI model since it maximizes company revenue
C) Reduce transparency in AI pricing decisions to avoid scrutiny
D) Expand the pricing model to cover more cities without modification
A) Implement ethical pricing caps to prevent exploitative surge pricing - AI pricing strategies should avoid unfairly burdening vulnerable populations.
What role does a notified body play in AI compliance under the EU AI Act?
A) Conducting conformity assessments for high-risk AI systems
B) Deploying AI models in regulated industries
C) Developing AI model documentation and audits
D) Training AI users on ethical guidelines
A) Conducting conformity assessments for high-risk AI systems - Notified bodies assess compliance with regulatory requirements before high-risk AI deployment.
A predictive policing AI system disproportionately increases patrols in low-income areas. What governance measure would be MOST effective in ensuring fairness?
A) Conduct an independent audit and adjust predictive models to prevent bias
B) Expand AI enforcement to more communities without reviewing bias
C) Reduce transparency about AI crime predictions to avoid challenges
D) Allow AI to make policing decisions without human intervention
A) Conduct an independent audit and adjust predictive models to prevent bias - AI-driven policing must be fair and avoid reinforcing societal inequalities.
A social media AI system amplifies sensational news stories for higher engagement. Regulators raise concerns about misinformation. What governance step should be prioritized?
A) Adjust ranking algorithms to prioritize factual content
B) Maintain current engagement-based ranking for maximum revenue
C) Reduce transparency in AI content moderation to avoid scrutiny
D) Expand AI-driven content recommendations without modifications
A) Adjust ranking algorithms to prioritize factual content - AI-driven content distribution should promote credible information.
A self-driving car AI system is programmed to prioritize passenger safety but has an increased accident rate for pedestrians. What governance measure should the manufacturer take?
A) Maintain current AI behavior since it optimizes passenger outcomes
B) Recalibrate AI decision-making to balance pedestrian and passenger safety
C) Reduce AI transparency to prevent legal issues
D) Automate all accident responses without regulatory oversight
B) Recalibrate AI decision-making to balance pedestrian and passenger safety - AI-driven transportation must adhere to ethical safety guidelines.
A medical AI system recommends treatment plans for patients but is found to be less effective for rare diseases due to limited training data. Which approach would MOST improve its performance?
A) Increase model complexity without adding new data
B) Collect more diverse medical data and retrain the model
C) Reduce AI involvement and rely solely on human doctors
D) Allow AI to continue making predictions without modifications
B) Collect more diverse medical data and retrain the model - AI in healthcare must be trained on representative datasets to ensure accuracy for all patient groups.
An AI-driven credit scoring system unintentionally denies loans to individuals in certain geographic areas due to historical data patterns. This is an example of:
A) Algorithmic transparency
B) Disparate impact
C) Model robustness
D) AI optimization
B) Disparate impact - AI decisions can unintentionally disadvantage certain groups even without explicit bias in programming.
Which AI governance challenge is MOST relevant for AI used in medical diagnoses?
A) Ensuring explainability and trust in AI-generated diagnoses
B) Reducing human involvement to optimize efficiency
C) Keeping medical AI models proprietary and confidential
D) Preventing regulatory oversight to speed up deployment
A) Ensuring explainability and trust in AI-generated diagnoses - AI in healthcare must be interpretable for effective patient treatment.
A self-driving car AI system must choose between two unavoidable accident scenarios. One option endangers the vehicle’s passenger, while the other risks harming pedestrians. Which AI governance principle is MOST relevant in designing the ethical response for this situation?
A) Fairness and bias mitigation
B) Explainability
C) Ethical decision-making frameworks
D) Data minimization
C) Ethical decision-making frameworks - Autonomous vehicle AI must incorporate ethical considerations to handle high-risk scenarios.
What is a key compliance requirement for high-risk AI systems under the EU AI Act?
A) They must undergo mandatory risk assessments and human oversight
B) They must be fully autonomous without human intervention
C) They must operate without transparency obligations
D) They are exempt from ethical review
B) They must undergo mandatory risk assessments and human oversight - The EU AI Act mandates rigorous compliance checks for high-risk AI applications.
A bank deploys an AI-powered fraud detection system provided by an external AI company. After deployment, the bank notices that the system incorrectly flags a large percentage of legitimate transactions as fraudulent. What governance action should the deployer take?
A) Work with the AI provider to recalibrate fraud detection criteria and report persistent risks
B) Maintain AI-driven fraud detection since it reduces financial risk
C) Reduce AI transparency to prevent scrutiny from regulators
D) Expand AI-based fraud detection without reviewing model fairness
C) Work with the AI provider to recalibrate fraud detection criteria and report persistent risks - Deployers must ensure AI models do not cause unjustified harm.
A ride-sharing platform’s AI surge pricing model dramatically increases fares during natural disasters. What governance response should the company take?
A) Introduce ethical pricing constraints during emergencies
B) Maintain AI-driven surge pricing since it follows market demand
C) Reduce transparency about AI pricing to prevent customer complaints
D) Expand AI surge pricing without regulatory intervention
A) Introduce ethical pricing constraints during emergencies - AI pricing models should not exploit users in critical situations.
A company uses AI to determine employee performance bonuses. Employees claim the system favors younger workers over older ones. What governance step should be taken?
A) Audit the AI system for potential age bias and adjust its decision criteria
B) Maintain the AI system since it optimizes employee productivity
C) Reduce AI transparency to prevent challenges from employees
D) Automate all performance decisions to remove human subjectivity
A) Audit the AI system for potential age bias and adjust its decision criteria - AI workplace applications must ensure equal treatment of employees.
Which aspect of the NIST AI RMF focuses on identifying AI risks and their impact?
A) Map
B) Measure
C) Govern
D) Manage
A) Map - The ‘Map’ function helps organizations identify AI risks and understand their implications.
What does the concept of AI fairness aim to prevent?
A) Bias and discrimination in AI decision-making
B) The use of AI in high-risk environments
C) The need for AI audits and compliance checks
D) The application of AI in public sector services
A) Bias and discrimination in AI decision-making - Fairness ensures AI models do not create or reinforce societal biases.
Which AI regulation primarily governs consumer privacy rights in the United States?
A) GDPR
B) CCPA
C) AI Fairness Act
D) HIPAA
C) CCPA - The California Consumer Privacy Act (CCPA) ensures consumer privacy protections in AI data processing.
Which of the following best defines artificial intelligence (AI)?
A) The ability of machines to perform tasks that typically require human intelligence
B) A set of algorithms that always function without human intervention
C) A type of software that replaces all human decision-making
D) A system that cannot learn from data and requires pre-programmed rules
A) The ability of machines to perform tasks that typically require human intelligence - AI encompasses machine learning, natural language processing, and computer vision.
An AI hiring tool consistently ranks male candidates higher for leadership roles. What is the best governance response?
A) Maintain current rankings as they align with historical hiring data
B) Reduce transparency in hiring decisions to prevent challenges
C) Audit the AI system for gender bias and adjust decision criteria
D) Expand AI automation in hiring without addressing bias concerns
C) Audit the AI system for gender bias and adjust decision criteria - AI hiring tools should ensure fair evaluation across gender identities.
Which aspect of AI security involves protecting AI systems from data poisoning?
A) Data integrity, which ensures training datasets are free from tampering
B) Model explainability, which ensures AI decisions can be interpreted
C) Algorithmic fairness, which ensures bias mitigation
D) AI automation, which prioritizes efficiency over security
C) Data integrity, which ensures training datasets are free from tampering - Data poisoning corrupts AI training data, leading to unreliable predictions.
Which principle ensures that AI decision-making processes can be understood by stakeholders?
A) Explainability
B) Automation
C) Algorithmic secrecy
D) Complexity
A) Explainability - This principle ensures transparency and accountability in AI decision-making.
What distinguishes AI from traditional software?
A) AI can learn and improve from data without explicit programming
B) AI always produces unbiased outcomes
C) AI does not require data for training
D) AI is only used for automation and does not involve decision-making
A) AI can learn and improve from data without explicit programming - AI differs from traditional software by utilizing machine learning and adaptive decision-making.
Which of the following is a major limitation of black-box AI models?
A) They lack transparency, making it difficult to understand how decisions are made
B) They are the most fair and unbiased AI systems
C) They require no human oversight
D) They always produce ethically sound outcomes
A) They lack transparency, making it difficult to understand how decisions are made - Black-box AI models make accountability and fairness evaluations challenging.
An AI model trained on past hiring data prefers candidates from elite universities. What governance measure should be implemented?
A) Conduct bias testing and diversify training data to ensure fairness
B) Maintain the hiring model as it reflects historical hiring success
C) Automate all hiring decisions to eliminate human subjectivity
D) Reduce transparency in hiring decisions to prevent regulatory challenges
A) Conduct bias testing and diversify training data to ensure fairness - AI recruitment systems should not reinforce exclusionary hiring practices.
