Brainscape's Knowledge GenomeTM

Browse over 1 million classes created by top students, professors, publishers, and experts.


See full index

AIGP > Module 8 > Flashcards

Module 8 Flashcards

1
Q

What is a tort?

A

A tort is a civil wrong that causes a claimant to suffer loss or harm, resulting in legal liability for the person who commits the tortious act

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What initiative in the EU proposes to reform tort law?

A

EU Commission’s AI Liability Directive

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is the objective of the EU Commission’s AI Liability Directive?

A
  • To provide similar compensation for AI-related claims as for damage incurred by other products
  • To address specific characteristics of AI such as opacity, autonomous behavior, complexity and limited predictability
  • To prevent companies from contracting away their liability for the products to which they contribute
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

What initiatives in the US propose to reform tort law?

A
  • State regulations on autonomous vehicles with specific safety standards
  • FTC guidance to businesses
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What guidance has the FTC provided to businesses to reform tort law and AI?

A
  • Be transparent with consumers about AI use and how algorithms work
  • Ensure decisions are fair, robust and empirically sound
  • Hold themselves accountable for compliance, ethics, fairness and non-discrimination
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Data licensing terms are used to regulate what concerns between parties? List 6 items

A
  • Designating certain model components as trade secrets
  • Protecting model components
  • Including assignment rights in model evolutions from one party or the other
  • Determining the license and use rights the parties want to establish between the provider and the user for each model component
  • Establishing rights in the terms and conditions
  • Liability and indemnification
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

How can you protect model components in data licensing terms?

A
  • Limiting the right to use the component(s)
  • Designating AI components as confidential information in the terms and conditions
  • Restricting the use of confidential information
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

List 3 typical exceptions to the IP infringement indemnity in traditional software or technology licensing agreements

A
  • Modifications to the software or technology
  • Unauthorized combination of the software or technology with other software or technology
  • Use of the software or technology beyond the scope authorized in the agreement
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Why are traditional exceptions to software or technology licensing agreements often a bad fit with AI?

A

Modifications and combinations often occur with the models, for example:
- Must be trained, which means modifications to the AI model solution
- Must be combined with training data and production data
- May evolve and exceed a pre-determined authorized scope over time

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

What should licensees insist on in relation to data licensing?

A
  • Minimum performance metrics to ensure that the licensed model provides adequate accuracy, reliability and robustness
  • Contractual warranties and indemnities from the licensor to mitigate the risk of underperformance
  • Thorough testing and validation, in advance of the license and on an ongoing basis
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

How can AI affect copyrights and patents?

A
  • If an AI system generates a work, it is unclear who owns the copyright–the developer of the AI system or the user
  • Similarly, if an AI system generates the basis for an invention, it raises questions about who should be listed as the inventor and who may file for the patent rights to that invention
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

What is the issue with AI and trademarks?

A
  • The use of AI systems to generate trademarks raises questions about the ownership and distinctiveness of such marks
  • For example, if an AI system creates a new logo or slogan, who owns the trademark and how can its distinctiveness be established?
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

What did the US Court of Appeals for the Federal Circuit recently determine in the case of Thaler v Vidal?

A

Only humans can be named as inventors on a patent

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Why did the European Patent Office (EPO) refused two patent applications designating AI as an inventor?

A

They determined that the inventory must have a legal personality

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What issues arise with the practice of data scraping or extracting data from publicly available sources or competitor websites?

A
  • Potential misuse of IP which can lead to IP infringement
  • Trade secret misappropriation
  • Unfair competition claims
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Provide 2 examples where raw data might be protected as IP

A
  • It meets specific national law conditions for qualifying as a protected work (a literary, artistic, or other type of work)
  • Is part of a database, as in most EU jurisdictions databases are protected by sui generis rights
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What does sui generis mean?

A
  • “of its own kind” - Latin
  • It is used in the legal context to describe cases, contracts, or legal rights that are unique and confined to their own facts, and therefore may not be of broader application
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

What should be included in AI literacy?

A

Both the technological and the human dimensions of AI; how it works (the techniques and the technologies) and what its impact is on people (e.g., privacy, agency, proprietary information, etc.)

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What is critical to include when training employees?

A

Ensuring they understand the purpose, limitations, and security and privacy controls for AI systems

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What should you include in your employee training program to address the risks specific to generative AI?

A

The importance of not providing sensitive, personal or classified information to an AI program without awareness of the required approval

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

Provide an example of the risk of not properly educating your employees

A
  • 2023, Samsung programmer
  • Uploaded a meeting recording to produce a presentation
  • Much of the information became public
  • Samsung cannot remove the information
  • The information may now be part of the ChatGPT training model
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

What does GPT mean?

A

Generative pre-trained transformer

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
23
Q

Which types of jobs are at a higher risk of automation than others?

A

Jobs that involve repetitive manual tasks

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
24
Q

According to Forbes, which sectors are most affected by AI automation?

A
  • Finance & banking
  • Media & marketing
  • Legal services
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
25
Q

Which jobs are less likely to be automated?

A

Jobs that rely heavily on skills like creativity, critical thinking and emotional intelligence

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
26
Q

What skills could individuals develop to better integrate AI into their jobs?

A
  • How to use AI
  • Prompt engineering
  • Data analysis
  • Coding
  • User experience (UX) design
27
Q

What 2 forms of human oversight exist in AI?

A
  • Oversight of the AI program itself
  • Bypassing or replacing the AI process with alternative processes
28
Q

Why is human oversight of the AI program important?

A
  • It is a best practice and in some cases a legal requirement
  • Critical to ensure the AI is performing as expected
29
Q

What is involved in human oversight of an AI program?

A
  • Generating risk assessments
  • Monitoring
  • Evaluating
  • Auditing
  • Testing the system on a regular basis
30
Q

Provide an example of a legal requirement in the US to provide an alternative process to bypass an AI system

A

US Customs and Border Protection processing for identity verification is required by law to offer an alternative to its primary biometric system of facial recognition

31
Q

List the current challenges in building a profession of certified third-party auditors globally, along with consistent frameworks and standards for them

A
  • There is no mature auditing framework in place detailing AI sub-processes
  • Auditors are challenged with how to perform audits successfully when there are no widely adopted precedents for handling AI use cases
  • Internal auditors, also known as first-party auditors, are under the auditing entity direct subordination - this can influence the scope of the internal audit, which cannot alone give rise to public accountability and might provide unverifiable assertions that the AI has passed legal or ethical standards
32
Q

What legislation has been proposed in the US to address AI auditing?

A

Algorithmic Accountability Act

33
Q

What important requirement is included in the US Algorithmic Accountability Act?

A

First-party audits

34
Q

What audit provisions are included in the GDPR?

A

The Federal Reserve and Office of the Comptroller of the Currency’s SR 11-7 guidance on model risk management suggests that an internal auditing team be different from the team developing or using the tool subject to audit

35
Q

List the possible solutions for building a profession of certified third-party auditors globally, along with consistent frameworks and standards for them

A
  • Leverage existing auditing frameworks and codes of ethics
  • Internal auditors need to build a new way of auditing AI
  • Make use of external audits which necessarily look backwards and will typically exhibit a range of independence from the deploying entity
36
Q

List 3 examples of existing auditing frameworks and code of ethics

A
  • COBIT Framework
  • Institute of Internal Auditors AI Auditing Framework
  • COSO ERM Framework
37
Q

What needs to be incorporated into existing audit practices to better cover the risks related to AI?

A

Data, models, outputs and processes to guarantee compliance, ethics and transparency

38
Q

What requirements are included in the EU Digital Services Act in relation to auditing?

A

It calls for third-party audits and takes the first steps towards defining independence for third-party auditors

39
Q

Who developed the COBIT 2019 framework?

A

Isaca

40
Q

How could the COBIT 2019 framework be leveraged as a starting point for auditing AI initiatives?

A

It provides auditors with tools, including process descriptions, desired outcomes, base practices and work products to enable the auditor to provide assurance over the AI initiative for any organization

41
Q

Why is the COBIT 2019 framework a good starting point?

A

It provides a good framework for considering the risk of any initiative or process within an organization

42
Q

What is the starting point for any audit within an organization?

A
  • Defining the scope and objectives of the audit
  • Considering the risks to the organization of each AI system
  • Compiling the risks into a document such as a risk and control matrix
43
Q

List 4 examples of risk related to AI strategy

A
  • Lack of alignment between IT plans and business needs
  • IT plans that are inconsistent with the organization’s expectations or requirements
  • Improper translation of IT tactical plans from the IT strategic plans
  • Ineffective governance structures failing to ensure accountability and responsibility for IT processes related to the AI function
44
Q

List 4 other AI auditing frameworks that have been published by international organizations and governments

A
  • Committee of Sponsoring Organizations (COSO) Enterprise Risk Management Framework
  • U.S. Government Accountability Office (GAO) AI Framework
  • Institute of Internal Auditors (IIA) Artificial Intelligence Auditing Framework
  • Singapore Personal Data Protection Commission (PDPC) Model AI Governance Framework
45
Q

Provide examples of codes of ethics for AI

A

UN Guiding Principles Reporting Framework
2019 Ethics Guidelines for Trustworthy AI (European Commission’s High-Level Expert Group on AI)

46
Q

Provide 2 examples of what AI auditing frameworks can be predicated upon

A
  • An audit of AI to ensure that the system(s) work without bias or discrimination (bias audits)
  • An audit of the rules, processes, frameworks and tools within an organization to ensure ethical and responsible development of AI
47
Q

Describe the auditing framework set forth by the Netherlands for government use of algorithms

A
  • First, it looks at governance and accountability
  • Second, it looks at models and data
  • Third, it looks at privacy
  • Fourth, it examines information technology general controls
48
Q

In the auditing framework set forth by the Netherlands for government use of algorithms, what do they include in Governance and accountability?

A
  • Management of the algorithm throughout its life
  • Who has what responsibilities
  • Where liability lies
49
Q

In the auditing framework set forth by the Netherlands for government use of algorithms, what do they include in Models and data?

A
  • Data quality
  • The development, use and maintenance of the model underlying the algorithm
  • Bias
  • Data minimization
  • Output testing
50
Q

In the auditing framework set forth by the Netherlands for government use of algorithms, what do they include in Privacy?

A

Compliance with GDPR

51
Q

In the auditing framework set forth by the Netherlands for government use of algorithms, what do they include in Information technology general controls?

A
  • Access rights to data and models
  • Security controls
  • Change management
52
Q

What conflicts in audit standards and methodology make audit results contestable or misleading?

A

Lack of agreed-upon methods by which an audit is conducted

53
Q

List 2 ways an audit is conducted

A
  • By what means it is conducted
  • By what standard it is conducted
54
Q

What 3 types of audits have been defined by UK regulators?

A
  • Technical audits that look under the hood at system components such as data and code
  • Empirical audits that measure the effects of an algorithmic system by examining inputs and outputs
  • Governance audits that assess the procedures around data use and decision architectures
55
Q

What was developed by the Ada Lovelace Institute?

A

A taxonomy of social media audit methods, focusing on scraping, accessing data through application programming interfaces and analyzing code

56
Q

Audits of automated decision systems, variously also called algorithmic or AI systems, are currently required by (list 3)

A
  • EU’s Digital Services Act
  • EU’s GDPR
  • Either required or considered in a host of U.S. laws
57
Q

What types of auditing can be key for process-based and outcome-based explanations?

A
  • Ex-ante auditing
  • Life-cycle auditing
58
Q

How can audits abet transparency and explainability?

A
  • They can make visible aspects of system construction and operation that would otherwise be hidden
  • They can also substitute for transparency and explainability - Instead of relying on those who develop and deploy algorithmic systems to explain or disclose, auditors investigate the systems themselves
59
Q

Why is AI automation crucial in keeping a competitive edge while meeting regulations?

A
  • AI governance can include manual data validation, comparison, and other intervention, which requires familiarity with the data management and handling process
  • When done manually, model validators may require expertise in each type of algorithm being used, which can be slow and costly to the business and can result in human errors
  • This delay in process could result in the company falling behind competitors or being late to provide information to auditors
60
Q

What is the advantage of automating AI governance?

A
  • Documentation and validation processes of AI governance would be more efficient
  • Institutionalizes processes, policies, and compliance of AI deployments to continuously collect evidence
  • Ensures consistency, accuracy, timeliness, efficiency, cost effectiveness and scalability of AI deployment
61
Q

What is AI Verify?

A
  • Launched by the Singapore government
  • An AI governance testing framework and toolkit to ensure systems meet declared performance benchmarks
62
Q

List the 11 AI ethics principles included in AI Verify

A
  1. Transparency
  2. Explainability
  3. Repeatability/reproducibility
  4. Safety
  5. Security
  6. Robustness
  7. Fairness
  8. Data governance
  9. Accountability
  10. Human agency and oversight
  11. Inclusive growth, societal and environmental well-being
63
Q

What is the Model Card Regulatory Check?

A

An app devoted to automating regulatory compliance of AI systems based on accepted documentation tools in the AI space

AIGP (9 decks)

Brainscape helps you reach your goals faster, through stronger study habits.
© 2024 Bold Learning Solutions. Terms and Conditions