Eight Countries and AI Companies Sign Paris AI Safety Declaration

On November 9, 2026, representatives of eight countries together with major AI companies including OpenAI, Anthropic, Microsoft, and Mistral AI signed the Paris Call for AI Safety and Security. According to The Decoder, the signatories committed to sharing evaluation procedures for high-performance models, disclosing known risks, supporting independent audits, and establishing shutdown mechanisms for the models under their control. This declaration brings together governments and leading AI companies, focusing on transparency in model safety evaluations, risk information disclosure, and the feasibility of independent audits. The "shutdown mechanism" clause requires participating companies to maintain safeguards that allow their models to be shut down. To date, the reported facts are limited to the signing date, the participants, and the commitments made; there have been no further reports on follow-up implementation arrangements or additional signatories.

AI overview of reporting · Updated 1 h前

Not enough continuous observations to show a trend yet.