Skip to content
uk.diplomatie.gouv.fr·Indexed by this site · Originally publishedDate not given· Governance →· Its place in the scan record →

16 governments and 10 AI companies sign the Paris Call for AI Safety

Paris Call for AI safety and security | France in the United Kingdom

AI summary

On November 9, 2026, 16 governments, 10 AI companies including Anthropic and OpenAI, and 7 civil society organizations signed the Paris Call for AI Safety, making three commitments on frontier model safety. First, establish a common framework for frontier model evaluation and risk management, focusing on dangerous misuse including cyber and CBRN, as well as risks from increased autonomy or misaligned models (including loss of control), and publish non-sensitive evaluation results in a standardized format and establish a public shared repository. Second, clarify that the whole ecosystem is accountable by role: developers commit to disclosing model capabilities, limitations and known risks, supporting qualified independent evaluations as appropriate, and implementing fail-safe mechanisms that can restrict, pause or shut down models during training and internal use. Third, establish information-sharing and early-warning mechanisms, with policymakers committing to safe and confidential channels for whistleblowers free from legal or professional retaliation. The statement remains open to new signatories.

Read originaluk.diplomatie.gouv.fr