Article
May 2026
Expertise and policy guidance to prevent major AI risks
CeSIA is an independent centre of expertise and think tank dedicated to preventing major risks associated with AI through its research and education activities, public policy analysis and strategic recommendations.
They trust us
Our work
Evaluating AI models for the European Commission
Selected to assess manipulation risks from frontier AI models.
Global Call for AI Red Lines
15 Nobel laureates. 300+ signatories. Launched at the UN General Assembly.
Advising French institutions on AI risks
We discuss with French institutions like the Senate, the Ministry of Defense, or INESIA, the French AISI.
About
A team of AI experts based in Paris. CeSIA works closely with the European Commission, the OECD and French institutions on the risks posed by frontier AI.
News
Article
Dec 2025
Why an International Agreement Establishing Red Lines for AI is Both Necessary and Realistic
Article
Dec 2025
The Global Call for AI Red Lines, Initiated by CeSIA, Is Launched at the UN
Press Release
CeSIA Condemns Commission’s Decision to Weaken the AI Act: “A Precedent the Industry Will Use to Undermine EU Law”
The essentials of AI safety, every two weeks
Every two weeks, CeSIA analyses the developments shaping the future of AI: technological breakthroughs, policy decisions and safety issues. An independent read on what's at stake.
15 October 2026
Digital slimemold
A cyberattack overheats French supermarket fridges, Anthropic warns of self-replicating AI, OpenAI fires three safety researchers, 400 scientists call for a global AI regulator.
Get in touch
Working on AI issues and want to connect with our team? Get in touch.