Macron calls for banning AI uses that threaten children as Paris AI safety appeal is signed

A call to limit risky AI for children

French President Emmanuel Macron said uses of artificial intelligence that could harm children's mental health and brain development should be prohibited. In an interview with the economic daily Les Echos, published online on the evening of 9 November 2026, he also called for safeguards to protect democratic and national security. He described the task as building a framework that accounts for three kinds of danger: children, democratic security and national security. The remarks came as the Paris Peace Forum prepared to open.

“If there is a risk for children from certain uses of AI, then it must be banned.”

Macron said companies would accept a pre-audit and pre-certification, with a prior study of their models before children can use them. He named OpenAI and Anthropic among the firms that would submit their models to independent evaluations and audits before making them available to minors. He added that Europe would push its own regulation and urged the Americans, the Canadians and others to follow, describing the start of a regulation for frontier AI.

The Paris appeal on AI safety

Nearly 50 political leaders, representatives of international bodies, technology companies and researchers signed the Paris Call for AI safety on 9 November 2026. Six countries signed: Bahrain, Canada, Finland, France, Norway and the Netherlands. Ten companies also signed, including OpenAI, Anthropic, Microsoft and Mistral AI, while civil society groups such as LawZero, SaferAI, AI Safety Connect and Ifri backed the text. The appeal sets three priorities: strengthening independent evaluations, improving the exchange of alerts about incidents, and defining responsibilities. It aims to promote a coordinated international response to the risks posed by the most advanced AI models.

Signatory countries6

Signatory companies10

Signatory countries

6

Signatory companies

10

Data from the article text

Oversight of AI assistants

An expert committee, the Conseil de l'IA et du numérique (CIANum), handed its report to Macron on 10 November 2026 during the Paris Peace Forum. Its central conclusion is that oversight of AI assistants, from ChatGPT to Gemini or Grok, should be tightened right away. The recommendations cover a field that has so far been less explored than the effects of social networks. Macron has also campaigned for banning certain social networks for children under 15. He framed the effort as a matter of structure rather than rejection.

“The idea is not to refuse these innovations but to create a framework that takes the dangers into account.”

Timeline and next steps

The sequence began on 9 November 2026 with the signing of the appeal and the publication of the Les Echos interview that evening. On 10 November the expert report reached Macron and the forum opened, with sessions scheduled for 10 and 11 November. The Élysée said the charter is meant to inform national and European legislative debates. Les Echos noted that the approach remains non-binding, even as the charter's signatories would commit to independent evaluations of generative AI systems and access to their models' data.

Limits of the approach

Macron said the current arrangement could not continue as it stands. He argued that the system should be changed, even up to a ban, if it proves harmful.

“This system is not sustainable, it must be changed up to and including banning it if it is bad.”

The commitments announced so far rely on voluntary participation rather than binding law. The Paris Peace Forum is scheduled to run on 10 and 11 November 2026.