AI ethics committee

An AI ethics committee, sometimes called an AI review board or responsible AI council, brings together perspectives a single function lacks: legal, privacy, security, risk, data science, the business and, ideally, people who understand affected communities. Its usual tasks are reviewing high-impact use cases, ruling on cases the policy does not settle, advising on risk acceptance, and overseeing that principles are applied in practice.

ISO/IEC 38507 reminds governing bodies that they remain accountable for the organization's use of AI, so the committee must have a clear charter: mandate, membership, decision rights (advisory or binding), escalation path to executives or the board, and how its decisions are recorded. The NIST AI RMF Govern function similarly expects defined roles and accountability. A committee without authority, resources or documented outcomes risks becoming symbolic, which regulators and auditors will notice.

Example: the committee of a health insurer blocks a proposal to use wearable data for premium decisions and asks for an impact assessment before any pilot.