What happened
In her State of the European Union address on 16 September 2026, European Commission President Ursula von der Leyen announced the forthcoming EU Kids Act, confirming it will bar children under 13 from social media accounts, restrict minors 13-15 to parent-supervised 'mini accounts,' and — significantly for AI governance — extend binding 'safe by design' obligations to AI chatbots and AI companion services (alongside social media, video-sharing platforms, and online games) that are accessible to minors. The full legislative text is scheduled for formal publication by the Commission on 17 September 2026, but the substantive content and the Commission President's public commitment to it occurred within the window via the SOTEU address and Commission briefings on 15-16 September. Age-verification, addictive-design bans, and safety-by-design mandates would apply to AI companion/chatbot providers serving EU users.
Why it matters
This is the first EU-level binding instrument to explicitly extend child-safety-by-design and age-assurance obligations to AI chatbots/companions as a named regulated service category, layering onto existing GDPR and AI Act transparency obligations (Article 50) for conversational AI. Any AI company offering companion/chatbot products accessible in the EU (OpenAI, Character.AI, Replika, Meta AI, etc.) would face new age-verification, default safety, and design-limitation requirements once enacted — a major expansion of AI-specific regulatory scope beyond the AI Act's existing risk-tiering.
Action needed
AI chatbot/companion providers serving EU users should monitor the formal EU Kids Act text (expected 17 September 2026) for age-verification and safe-by-design obligations, and begin scoping compliance with anticipated minor-protection design mandates ahead of the co-legislative process (European Parliament and Council negotiation).