Principles, Declarations & Company Policies
11 entriesStatements of values from governments, researchers and AI developers.
Foundations, Papers & Books
10 entriesCore academic and popular reading on AI ethics and fairness.
Fairness, Bias & Responsible AI Tools
7 entriesSoftware for measuring and reducing bias and documenting models.
Research Institutes, Advocacy & Safety Groups
15 entriesOrganizations working on AI ethics, society and safety.
Incidents, Risk Trackers & Trends
9 entriesEvidence on what goes wrong and how AI is spreading.
Courses & Learning
4 entriesFree introductions for learners.
Related Technologies on This Site
11 entriesGovernance, risk, explainability and privacy pages on this site.
Frequently Asked Questions
What is AI ethics?
AI ethics is the study and practice of making sure AI systems are built and used in ways that respect human rights and values. Common themes are fairness and non-discrimination, transparency and explainability, privacy, accountability, safety and human oversight.
Which principles are most widely cited?
The OECD AI Principles, the UNESCO Recommendation on the Ethics of AI and the European Commission's Ethics Guidelines for Trustworthy AI are among the most widely referenced. A 2019 study in Nature Machine Intelligence compared 84 sets of guidelines and found common themes such as transparency, justice and fairness, non-maleficence, responsibility and privacy.
How do ethics, governance and regulation differ?
Ethics sets out values and principles, governance turns them into policies, roles and controls inside an organization, and regulation makes some requirements legally binding, as the EU AI Act does. See the AI Governance Frameworks page on this site for the regulatory and standards layers.
How can bias be measured and reduced?
Toolkits such as AI Fairness 360, Fairlearn and Microsoft's Responsible AI Toolbox provide fairness metrics and mitigation methods. Documentation practices such as model cards and datasheets for datasets help others judge whether a model or dataset suits a use. Research also shows that some fairness criteria cannot all be satisfied at once, so trade-offs need to be chosen deliberately.
Where can I track real-world AI harms?
The AI Incident Database and the OECD AI Incidents Monitor collect reported incidents, the MIT AI Risk Repository catalogues risk types, and the Stanford AI Index reports yearly trends including responsible AI.