In brief
French experts issued 18 recommendations after a study highlighted that 33 million people used generative AI and a quarter worry about its risks. They call for mandatory risk assessments before launch, design safeguards against flattery and anthropomorphism, early‑stage AI literacy, and an inter‑ministerial protocol to stop disinformation. The report stresses the tools could weaken cognition, increase emotional attachment and inflame online violence.
On 10 November 2026, a French commission of experts announced a set of 18 recommendations that call for tighter rules on generative artificial intelligence, citing risks to people’s cognitive function, mental health and social cohesion.
The report, titled Protecting Without Hindering: Vulnerabilities in the Age of Generative AI, notes that 33 million unique users in France had adopted generative AI by the start of the year and that a quarter of the population expressed concerns about its potential dangers.
Key recommendations include a mandatory multidisciplinary risk assessment before any consumer generative‑AI system can enter the market, strict design safeguards aimed at eliminating excessive flattery and anthropomorphism so that users’ psychological vulnerabilities cannot be exploited for commercial gain, and large‑scale AI education beginning in early childhood to foster critical thinking.
The report also warns that generative AI could amplify disinformation, threaten democracy and social cohesion, and that it can "weaken users' cognitive abilities, foster emotional attachment to these tools and amplify online violence."
"Excessive use of AI could weaken users' cognitive abilities, foster emotional attachment to these tools and amplify online violence."
To counter such risks, the commission recommends an inter‑ministerial protocol that can be activated quickly to detect and respond to AI‑related misinformation.
Sources: Xinhua (11 November 2026) and the commission’s own research findings.