EU Proposes AI Kids Act to Regulate Child‑Focused Chatbots

The European Union is advancing a legislative draft that would set tight restrictions on chat‑based artificial‑intelligence tools aimed at youngsters, naming well‑known services such as OpenAI's ChatGPT, Character.ai and Snapchat's My AI.

Called the AI Kids Act, the proposal intends to remove from minor‑targeted conversational agents any features regulators deem overly persuasive or exploitative. Under the text, a chatbot marketed to anyone under 18 would be prohibited from using personalized suggestions, gamified reward mechanisms, or similar design tricks that encourage extended use. It also demands strong age‑verification systems and requires that all data gathered from minors be erased after a brief retention period.

The measure expands on the EU's wider AI Act, which categorises high‑risk AI systems and sets compliance duties across industries. Lawmakers contend that chat agents pose a distinct danger because they combine natural‑language dialogue with adaptive learning, fostering a sense of companionship that can shape young users' attitudes and behaviour. By limiting these "affective" elements, the bloc aims to safeguard children’s privacy and mental health while still permitting educational content.

Industry bodies have voiced worries about how the draft would work in practice. Developers point out that many of the targeted capabilities—such as personalised tutoring or context‑aware help—are central to the tools' educational value. They caution that sweeping bans could choke innovation and drive minors toward unregulated, possibly unsafe alternatives. The European Commission, however, argues the steps are proportionate and that compliance routes, including optional certification, will be provided.

Public response is mixed. Consumer‑rights groups applaud the focus on youths, citing research that links heavy screen time and algorithmic persuasion to attention‑deficit issues in young people. Opponents claim the regulation is paternalistic and that parental controls could achieve comparable results without industry‑wide mandates. The discussion mirrors a broader global debate on balancing AI‑driven learning benefits with the need to protect vulnerable users.

The AI Kids Act is now set for consideration by the European Parliament and the Council of the European Union. Should it be adopted, it may become a model for other regions grappling with the fast‑moving uptake of generative AI in daily life. Firms operating in the EU will likely have to redesign interfaces, revise data‑handling practices, and perhaps launch separate, trimmed‑down versions of their products for younger audiences. The exact wording and timetable remain unclear, but the proposal marks a decisive move toward tighter oversight of AI technologies that interact directly with children.