CeSIA

CeSIA (Centre pour la Sécurité de l'IA), the French Center for AI Safety, is an independent non-profit based in Paris working to prevent catastrophic risks from frontier AI. Launched in May 2024, CeSIA now operates through three pillars: French policy, with France's 2026 G7 presidency as an active lever; EU AI Act enforcement; and international coalition-building; backed by public advocacy through op-eds, media partnerships, and content creators.

CeSIA initiated and coordinates the Global Call for AI Red Lines together with The Future Society and the Center for Human-Compatible AI. Launched at the UN General Assembly by Nobel laureate Maria Ressa in September 2025, the campaign has been endorsed by 12 Nobel Prize recipients, 11 former heads of state, and 90+ organisations, and has generated over 300 media mentions. Follow-on work includes six workshops with representatives from 20+ countries, a forthcoming OECD-commissioned report, and a standing UNESCO working group.

On policy, CeSIA was a direct interlocutor and advisor to the OECD during the final phase of the Hiroshima Reporting Framework, where many substantive recommendations, including risk thresholds, unacceptable risk definitions, and whistleblower protection, were adopted verbatim; the OECD told CeSIA it was essential to the success of the process. CeSIA has also had recommendations adopted into the EU AI Act's Code of Practice. CeSIA testified before France's National Assembly and has been invited to testify before the European Parliament's Committee on Security and Defence.

The EU AI Office selected CeSIA, ahead of larger competitors including Deloitte and KPMG, to design and build evaluations for harmful manipulation under the AI Act, working in a consortium with Transluce and EquiStamp. CeSIA also teaches France's only university-level AI safety courses, at ENS and Sciences Po, and authored the AI Safety Atlas, an open textbook used by 2,000+ students and recommended by 80,000 Hours.