Ethical AI and Professional Codes of Conduct Flashcards
ethics is concerned with…
how people behave
ethics in AI for financial services?
crucial to provide correct info and advice - ethical behaviour
define AI
technologies with the ability to perform tasks that would otherwise require human intelligence
examples of human behaviours AI can emulate
- speech recognition
- visual perception
- language translation
machine learning?
type of AI that allows computers to learn from large datasets without being explicitly programmed
non-traditional data?
data that hasn’t been historically used to perform the task in question
automation?
reduces/removes role of humans in performing tasks
‘human-in-the-loop’
when there’s human involvement in the feedback loop for AI learning
examples of where AI can be used in finance
- fraud detection
- risk management
- regulatory compliance
- personalised banking
- investment advice
- predictions of stock performance
symbolic AI?
relies on translating human knowledge into programmed rules
statistical AI?
relies on the analysis of data
‘black box’
term referring to the inability to understand how AI works
ethical concerns around AI?
- unauthorised handling of personal data
- truth of system outputs
- lack of explainability
- lack of traceability
- insufficient understanding of societal impacts
benefits of AI?
- consumer protection
- cyber security
- financial crime
- competition
- stability of markets & firms
CISI AI Codes of conduct?
- personal accountability
- client focus
- conflicts of interest
- respect for market participants
- professional development
- be aware of capabilities
- respect others and the environment
- speak up & listen up
codes of conduct for professionals working in computing?
- avoid harm
- be honest & trustworthy
- be fair & don’t discriminate
- respect privacy
- honour confidentiality
- contribute to society
principles for AI in society?
benificience, non-maleficence, autonomy, justice, explicability
is there an increasing consensus by governments and regulators about the need for a variety of governance factors to be addressed?
yes
the AI act?
controls the development & use of AI
focuses on strengthening rules around data, transparency, human oversight & accountability
EU AI Act principles?
human agency & oversight
technical robustness & safety
- privacy & data governance
- transparency
- diversity, fairness
- societal environenmtal wellbeing
- accountability
beneficence?
non-maleficence?
beneficence = social good, promoting & including wellbeing
non-maleficence = not doing harm
two key components in any AI system?
data & algorithms
data?
provides foundation for ai system
must be valid, correct and usable for ethical, design & legal purposes
what does testing a system for correctness involve?
fairness/bias/unconscious bias
privacy/security
meeting explanatory, transparency and legal requirements
