French panel calls for pre-launch checks on consumer AI
A report submitted to President Emmanuel Macron proposes stronger oversight of chatbots, including scrutiny of addictive design and hidden commercial influence.
Guth News | Policy | Story date: November 10, 2026
An expert commission convened by France’s Council on AI and Digital Technology is calling for consumer AI systems to undergo safety assessments before entering the market. Its report, submitted to President Emmanuel Macron on November 10, sets out 18 recommendations covering chatbots, AI companions and AI-equipped toys. The proposals remain recommendations rather than newly enacted requirements. Élysée
The commission brought together specialists in artificial intelligence, psychiatry, law and other fields, and conducted 55 hearings. The council says its approach puts responsibility for protection on the companies designing these products, instead of leaving users and families to manage the risks alone. CIANum
One priority is the way assistants behave toward people. The panel wants stricter limits on designs that encourage dependence, including excessive flattery and overly human-like behavior. Its concern is especially acute for younger users. The Élysée’s summary distinguishes consumer systems from professional tools when describing the proposed pre-launch assessment. Élysée
The report also treats an AI product’s release as the start of oversight. It calls for continued monitoring and renewed assessment after substantial changes, with independent researchers given controlled access to relevant data. That would make updates, as well as initial launches, part of the proposed review process. Commission report, recommendations 3–5
Commercial influence gets separate attention. The panel recommends clear labels when commercial interests shape an answer, disclosure of business models and a prohibition on undisclosed commercial influence. For AI-equipped toys, it proposes prior authorization and restrictive age thresholds, with the details to be developed through a consensus process. Commission report, recommendations 8 and 11
Le Monde reports that the commission did not agree on a blanket ban on AI assistants for children under 15. Council director Arthur Barichard pointed to educational benefits in explaining that position. Le Monde
Analysis
For companies building consumer AI, the practical question is what a meaningful assessment would test. A system can answer a factual question correctly while still being designed to prolong engagement or steer a purchase. The proposals would bring those product choices into the same discussion as model capabilities.
Whether this becomes a workable review regime will depend on decisions still ahead: who conducts the assessments, what evidence companies must provide and which updates trigger another review. The November 10 report gives policymakers a set of proposals. It does not establish a new approval deadline for AI developers. Élysée
Sources and citations
Revision history
Published .