Definition · Plain-language
AI ethics
AI ethics is the set of principles guiding the responsible design and use of AI, including fairness, accountability, transparency and human oversight.
The step most authors miss
Doing CRediT right? Don’t stop at the statement.
A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.
Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.
The core principles
AI ethics is usually expressed as a set of principles that recur across major statements. Fairness asks that systems do not produce unjustified, discriminatory outcomes. Accountability requires that responsibility for AI decisions can be located with people and organisations. Transparency and explainability ask that how a system works and decides can be understood and scrutinised. Privacy protects individuals’ data. Safety and security guard against harm and misuse. Human oversight ensures meaningful human control over consequential decisions. Underpinning these is respect for human rights, dignity and societal wellbeing.
Key international statements
Several authoritative statements give these principles weight. The OECD AI Principles, adopted in 2019 and updated since, were the first intergovernmental standard on AI and stress inclusive growth, human-centred values, transparency, robustness and accountability. The UNESCO Recommendation on the Ethics of AI, adopted by all member states in 2021, is the first global standard-setting instrument on AI ethics, grounding it in human rights and setting out concrete policy areas. These instruments are not binding law, but they shape national policy, regulation and the design of governance frameworks worldwide.
From ethics to practice
Principles alone do not change behaviour; AI ethics becomes effective only when operationalised. This is the bridge to responsible AI and AI governance: ethical principles supply the "what" and "why", while governance frameworks supply the "how" — translating fairness into bias testing, accountability into defined roles, transparency into documentation and oversight into review gates. The risk to guard against is "ethics washing", where principles are stated but never implemented. Embedding ethics therefore means tying each principle to measurable controls, assurance such as audit, and clear lines of responsibility.
Key facts
At a glance
- Definition: principles for the responsible design and use of AI
- Core principles: fairness, accountability, transparency, privacy, safety, human oversight
- Foundation: respect for human rights and human dignity
- OECD AI Principles: first intergovernmental AI standard (2019)
- UNESCO Recommendation: first global AI-ethics instrument (2021)
- Relation: ethics supplies values; governance operationalises them
Common misconceptions
What people often get wrong
Often heard: AI ethics is just abstract philosophy with no practical effect.
Actually: Ethical principles are increasingly operationalised through governance frameworks, audits and regulation, turning values such as fairness and accountability into measurable controls. They shape national policy and standards worldwide.
Often heard: AI ethics and AI governance are interchangeable terms.
Actually: AI ethics is the set of principles (the values); AI governance is the machinery of frameworks, roles and processes that enforces them. Ethics defines what is right; governance makes it happen.
Often heard: Publishing a set of AI principles means a company is ethical.
Actually: Stated principles without implementation are sometimes called "ethics washing". Genuine ethical practice requires tying principles to controls, assurance and accountability, not just a published statement.
Going deeper
