UNESCO Recommendation on the Ethics of AI ME: paras 131 to 134 Monitor and evaluate AI ethics policies credibly and transparently, with broad participation, indicators and mechanisms such as ethics commissions and observatories

Member States should credibly and transparently monitor and evaluate AI ethics policies, programmes and mechanisms using quantitative and qualitative approaches, with UNESCO developing the Ethical Impact Assessment and readiness assessment methodologies, ex ante and ex post policy evaluation, and evidence sharing. Monitoring should include vulnerable people and ensure social, cultural and gender diversity; tools and indicators should measure effectiveness against targets, including for disadvantaged groups, continuously and proportionately to risk, including evaluations and self-evaluations of institutions and providers. Possible mechanisms include an ethics commission, AI ethics observatory, repository, experience sharing, regulatory sandboxes and an assessment guide for AI actors.