EU Drafts AI Kids Act Targeting Chatbots Popular With Minors

The European Union is moving forward with a legislative proposal that would impose strict limits on chat‑based artificial intelligence tools aimed at children, including high‑profile services such as OpenAI's ChatGPT, Character.ai and Snapchat's My AI.

Named the AI Kids Act, the draft regulation seeks to strip minor‑focused conversational agents of features that regulators say can be overly persuasive or exploitative. Under the proposal, any chatbot that is marketed to users under 18 would be barred from employing personalized recommendations, gamified reward systems, or other design elements that encourage prolonged interaction. The text also calls for robust age‑verification mechanisms and mandates that all data collected from minors be deleted after a short retention period.

The initiative builds on the EU's broader AI Act, which classifies high‑risk AI systems and imposes compliance obligations across sectors. Lawmakers argue that conversational agents present a unique risk because they blend natural‑language interaction with adaptive learning, creating a sense of companionship that can influence young users' attitudes and behavior. By curbing these "affective" features, the EU hopes to protect children’s privacy and mental wellbeing while still allowing access to educational content.

Industry groups have raised concerns about the practical impact of the proposal. Developers note that many of the targeted functionalities—such as personalized tutoring or context‑aware assistance—are integral to the educational value of the tools. They warn that overly broad restrictions could stifle innovation and push younger users toward unregulated, potentially unsafe alternatives. The European Commission, however, maintains that the measures are proportionate and that compliance pathways, including optional certification, will be offered.

Public reaction has been mixed. Consumer‑rights advocates welcome the focus on minors, pointing to studies that link excessive screen time and algorithmic persuasion to attention‑deficit issues among youth. Critics argue that the regulation may be paternalistic and that parental controls could achieve similar outcomes without sweeping industry mandates. The debate reflects a wider global conversation about how to balance the benefits of AI‑driven education with the need to safeguard vulnerable users.

The AI Kids Act is now slated for review by the European Parliament and the Council of the European Union. If adopted, it could set a precedent for other jurisdictions grappling with the rapid adoption of generative AI in everyday life. Companies operating in the EU will likely need to redesign user interfaces, adjust data‑handling practices, and possibly roll out separate, stripped‑down versions of their products for younger audiences. The final shape of the law, and its timeline, remain uncertain, but the proposal signals a decisive step toward more stringent oversight of AI technologies that interact directly with children.