Anthropic, OpenAI, Microsoft and Black Forest Labs back a French-led charter on safer AI for children

Élysée · Declaration

Read the source (opens in a new tab)
What happened
The Élysée published the Charter of Engagement for Safer AI for Children on November 10, 2026, the industry partners' contribution to a children's AI coalition led by France, Spain and Kenya with ten UN bodies.
Key points
  • Participating companies whose AI systems minors can access will test them for risks to children before deployment where relevant and at regular intervals, and detect and respond to serious incidents involving children.
  • The coalition will set up a research network under an independent scientific committee; companies contribute funding, expertise and, where the law allows, access to models and data.
  • Results, including null or negative ones, will be published; companies may audit evaluation methods but gain no power to delay or block findings, and industry money funds only the shared-measures research strand.
  • Its principles include safety by design, age verification that does not rely only on self-declaration, and a ban on using AI to produce child sexual abuse material.
Evidence
A voluntary charter published in French by the French presidency; the summary here is a translation. Signatories are encouraged to report progress yearly through existing mechanisms, and the text names no enforcement mechanism.
Uncertain
The research network, its scientific committee and the independent coordinating organization are still to be set up, and the charter gives no funding amounts or timetable.

See it in the latest record