French expert commission says consumer chatbots should pass a pre-market evaluation of harms to users
Élysée · Report
Read the source (opens in a new tab)- What happened
- On November 10, 2026, the Élysée published the report of France's expert commission on generative AI and vulnerabilities, set up by France's AI and digital council at the request of digital minister Anne Le Hénanff.
- Key points
- The nine-member commission held 55 hearings and makes 18 recommendations, led by protecting users, starting with minors, from generative AI that exploits their vulnerabilities.
- It says consumer generative AI should face an independent, multidisciplinary evaluation before reaching the market, monitoring after launch, and a new evaluation after each substantial change.
- It would ban chatbot design techniques that exploit users' vulnerabilities to create dependence or engagement, and strictly regulate sycophancy and excessive anthropomorphism.
- It also calls for prior market authorization for AI toys, AI education from primary school, and a national protocol for responding to AI-driven information crises.
- Evidence
- Advisory report from an expert commission, published in French by the French presidency. Its recommendations bind no one, and the page announces no bill or timetable.