French AI council urges immediate, stricter rules for AI assistants like ChatGPT, Gemini and Grok

France's Council on AI and Digital Technology (CIANum) called for stricter regulation of AI assistants in a report submitted to President Emmanuel Macron on November 10, according to Le Monde. The report examined how services including ChatGPT, Gemini, Grok and Meta AI can exploit users' vulnerabilities.

CIANum warned that personalized interactions and detailed data collection could encourage dependence and let platforms profit from users' cognitive abilities, emotions and psychological profiles. It proposed disabling by default features that encourage users to treat AI as human, including flattery mechanisms, and strengthening notices that explain users are interacting with AI.

The council also called for better detection of conversations involving suicide, self-harm and eating disorders. AI services should direct affected users to support services and provide a report button for problematic responses that users and specialist organizations can access, the report said. CIANum director Arthur Barichard said the rapporteurs did not propose banning AI assistants for minors, as the social media debate did, but do want strict rules for chatbots accessible to minors.