The proposal presented by the European Commission on 17 September would also prohibit chatbots from simulating an emotionally dependent relationship among those under 18.
On 17 September 2026 the European Commission adopted its proposal for the EU KIDS Act, a text on the safety of online minors that explicitly encompasses conversational systems. IA chatbots and companions accessible to minors should be disabled by default, and they would be prohibited from simulating an interpersonal relationship that would create emotional dependence in a user under the age of 18.
Age levels and reverse burden of proof
The text organizes the access in tranches: no social networks before 13 years, accounts limited to one hour a day and under parental supervision between 13 and 15 years old, autonomous management of the account from 15 years old. Added to this are the framework for infinite scrolling, the removal of reward mechanics, the prohibition of nocturnal notifications and private default profiles. Age verification would go through privacy-friendly tools, including a European application that does not retain identity documents or biometric data.
Change of method is at least as important as measurement: it would be up to the very large platforms to demonstrate that their service is safe for children, rather than the regulators to prove the damage. For generalist companions and chatbots, application would fall under the European AI Regulation. The proposal is now in the hands of Parliament and the Council, and its entry into force is not expected until 2028.
Read the full article on the original source
Go to the source →