Responsible AI
Definition
An approach to designing, developing and using AI so that it is lawful, safe, fair, transparent, accountable and respectful of human rights throughout its lifecycle.
Responsible AI is the set of values an organization commits to when it builds or uses AI, and the practices that make those values verifiable. The OECD AI Principles, the UNESCO Recommendation on the Ethics of AI and the trustworthiness characteristics of the NIST AI RMF converge on similar themes: respect for human rights and the rule of law, fairness and non-discrimination, transparency and explainability, robustness and safety, privacy, and clear accountability.
The term is broader than legal compliance. A system can meet every applicable rule and still produce outcomes that users find unfair or opaque. Responsible AI therefore relies on governance to assign decisions, on impact and risk assessments to anticipate harm, and on human oversight and monitoring to correct it. For example, a team launching a recruitment screening tool would test it for bias across groups, explain its role to candidates and keep a human reviewer for rejections. These foundations are covered in RAIF.
Where this term is assessed
Assessed in the Responsible AI Foundation (RAIF), domain 3: Principles and frameworks.
Also covered in: Responsible AI Practitioner (RAIP)
Primary sources
- OECD AI Principles
- NIST AI RMF 1.0
- UNESCO Recommendation on the Ethics of AI