TL;DR
A French expert commission delivered 18 recommendations on AI assistants to President Emmanuel Macron’s office on November 10, including pre-launch evaluations and recurring checks for consumer chatbots. The proposals are recommendations, not law, and many would require action at European Union level.
A French expert commission delivered a report to President Emmanuel Macron’s office on November 10, recommending that consumer AI chatbots and companions pass an evaluation before entering the European market and face regular checks after launch. The 18 proposals, nine designated as priorities, are intended to protect users from systems that exploit or worsen vulnerabilities; they are recommendations, not current law.
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
The report, titled Protect without preventing: vulnerabilities in the age of generative AI, says evaluations should establish that consumer systems “neither exploit nor aggravate the vulnerabilities of their users.” It includes chatbots built into social media platforms. The commission suggests repeat assessments every 12 to 24 months, for example, as well as checks after substantial model changes or alerts. If a provider fails an evaluation, the report says withdrawing the system from the market “could be considered.”
The recommendations also target product design and advertising. The commission proposes restricting techniques that exploit vulnerability to encourage engagement or dependence, naming flattery and features that make AI seem human. It calls for a ban on disguised advertising in AI responses and proposes giving users rights over information chatbots infer about them.
For services open to minors, providers and deployers would either apply protective measures to everyone or use what the report calls robust age verification. Suggested safeguards include a button to end a conversation immediately, default time limits for minors, no persistent conversation memory by default, and human-like features switched off unless a parent agrees. The commission also proposes prior approval for AI toys. These are proposed measures; the report does not itself impose them.
How Pre-Launch Checks Could Change Access
Pre-launch assessment would change when authorities examine consumer AI systems. Under the approach described by the commission, scrutiny would happen before a product reaches the European market, then recur at set intervals or after specified developments. That could make safety assessments an ongoing part of market access, rather than a matter considered only after a problem is reported.
The recommendations focus on risks that may be difficult for users to identify or manage on their own, including emotional dependence, exposure of minors, and hidden commercial influence. The commission argues that interface choices and business models can shape how people use chatbots. If adopted, measures such as limits on persistent memory for minors or clear separation of advertising could affect how services are designed and how users interact with them.
The potential reach is substantial, but the report does not establish how many products would fail an evaluation, what the costs would be, or how enforcement would work. Those questions would depend on the legal rules eventually negotiated. The proposals also cover consumer systems; the Élysée says professional systems are outside the core proposal.
Products Worth Considering
Apple iPad 11-inch (A16), 128GB, Wi-Fi, Silver
Samsung Galaxy Tab A11 8.7-Inch Wi-Fi Tablet, 64GB, 4GB RAM, 90Hz, Gray, with 20W Charger
Samsung Odyssey G3 G30D 27-Inch FHD Gaming Monitor
As an affiliate, we earn on qualifying purchases.
Commission’s Mandate and Evidence
The commission, named “Generative AI and Vulnerabilities,” was set up by France’s advisory Council on Artificial Intelligence and Digital Technology, known as CIANum. According to the Élysée, Minister Delegate for AI and Digital Affairs Anne Le Hénanff assigned the work on February 20, 2026. Nine experts from different fields took part, led by AI researcher Serena Villata and psychiatrists Raphaël Gaillard and Amine Benyamina. The report says the group held 55 hearings over four months with health professionals, researchers, companies, regulators, and child-protection groups.
The commission cites figures to describe use and trust in France. Its report says 26% of 12- to 17-year-olds use generative AI daily, citing France’s 2026 digital barometer. It also cites a YouGov survey for Interface in which 92% of teenagers said they spend more time on generative AI apps than they would like. The Élysée summary says there were 33 million unique generative AI users in France at the start of the year. These figures come from different sources and refer to their respective survey populations and periods; they do not, by themselves, establish that chatbots caused harm.
The report groups its concerns around people handing over thinking or decisions to AI, design that may encourage emotional dependence, and the use of generative AI to facilitate online violence, including sexual deepfakes. It also raises broader concerns about public information, elections, and a possible “economy of cognition” in which companies profit from users’ vulnerabilities. These are the commission’s risk assessments, not findings that every chatbot or provider engages in such conduct.
““a paradigm shift””
— The Élysée
From Recommendations to EU Rules
The commission’s report does not change the law. Many of its recommendations would require measures at European Union level, and the report is framed as input to France’s position on updating the EU AI Act and other texts still under negotiation. As of November 11, the source material identifies no bill or timetable for turning the proposals into rules.
It is also not clear which recommendations the French government will support, how an evaluation would be run, or which authority would supervise it. The report gives a possible interval of 12 to 24 months for recurring checks, but does not establish a final schedule. The criteria for assessing vulnerability, the treatment of different types of systems, and the consequences of a failed assessment remain matters for policymakers.
Macron said in an interview that risky AI uses for children should be banned. That statement signals concern about minors’ safety, but the source material does not specify which uses he meant or establish that the commission’s proposals have been adopted as government policy.
French Position in EU Talks
The next step is for French officials to consider the report as they shape the country’s position on EU AI policy. Any binding pre-launch assessment or recurring monitoring requirement would depend on future legislative and regulatory decisions, including negotiations on the EU AI Act and other European texts. The report itself does not set a date for those decisions.
Further developments to watch include whether France formally advances any of the 18 recommendations, how EU institutions and other member states respond, and whether proposals on minors, advertising, user data, or AI toys are taken forward. Until authorities publish a concrete proposal, the scope and timing of any new obligations remain undetermined.
Key Questions
Has France made pre-launch chatbot testing law?
No. The idea is one of 18 recommendations in a report delivered to the Élysée on November 10. The report is not legislation.
Which AI products would the proposed checks cover?
The central proposal concerns consumer chatbots and AI companions, including chatbots built into social media platforms. The Élysée says professional systems are outside the core proposal.
How often could chatbots be checked after launch?
The report gives every 12 to 24 months as an example. It also proposes checks after a substantial model change or when someone raises an alert; no final schedule has been set.
What safeguards does the report propose for minors?
Options include protective measures for all users or robust age verification. Suggested measures include default time limits for minors, no persistent conversation memory by default, an immediate stop button, and human-like features switched off unless a parent agrees.
When could the recommendations take effect?
No implementation date is given. Many proposals would require EU-level action, and as of November 11 the source material identifies no bill or timetable.
Source: Thorsten Meyer AI
