French Expert Commission Presents 18 Recommendations to Regulate AI Chatbots

An expert commission delivered a report to President Emmanuel Macron on Tuesday, November 10, 2026, with 18 recommendations to better protect users of generative AI, starting with minors. Its recommendations include strict limits on design features that exploit users’ vulnerabilities and a requirement that consumer AI systems pass an evaluation before they reach the market. The Élysée’s statement does not announce a bill.

Who Wrote the Report

The report, titled “Protéger sans empêcher” (Protect Without Preventing), was written by the commission “IA générative et vulnérabilités.” The Conseil de l’IA et du numérique, an advisory body, set it up at the government’s request, and it has nine independent experts. They are led by AI researcher Serena Villata and psychiatrists Raphaël Gaillard and Amine Benyamina. The commission held 55 hearings with health professionals, researchers, companies and child protection groups, according to the Élysée.

What It Recommends

The commission recommends strictly regulating excessive flattery and anthropomorphism, meaning chatbots that behave as if they were human, directly in the design of generative AI systems. It also says consumer systems should undergo a multidisciplinary evaluation before being released, rather than being checked only after problems appear. It recommends training people in AI from an early age, and a clear, fast protocol for the state to detect and respond to information manipulation.

Other recommendations reported by AFP and other outlets include independent evaluations before and after launch, access to training and usage data for authorities and accredited researchers, and tougher enforcement against AI-generated non-consensual sexual content and child sexual abuse material. For minors, the report recommends high safety standards, including filtering and blocking dangerous requests. The Élysée said hidden advertising in chatbot answers should be banned.

The Data Behind the Report

The report says generative AI had 33 million unique users in France at the start of the year, and that a quarter of French people express concern about its risks. It also says 73% of French people use AI to access information. Nearly 70% of 15-to-24-year-olds use it regularly, AFP reported. The experts’ main worries are cognitive weakening, emotional attachment to chatbots and the amplification of digital violence such as sexual deepfakes.

France has been active on AI governance. In September, it convened a UN Security Council session on the subject, as we reported.

Read also: German companies unprepared for AI-based cyber attacks, and Fears over AI market slow impact global tech stocks.

Source: Élysée press release, November 10, 2026; AFP via L’Info durable; Presse-citron; Technomedia. Published November 10, 2026.

Photo by Tim Witzdam on Pexels