The European Commission’s proposed EU KIDS Act aims to regulate AI companions accessible to minors, focusing not only on harmful outputs but also on interaction designs that risk fostering emotional dependency in children and adolescents.
- New EU rules focus on preventing emotional dependency on AI companions for children.
- Providers must avoid design features that simulate relationships likely to foster dependency.
- Regulations apply to general conversational chatbots as well as dedicated AI companions.
What happened
On September 17, the European Commission put forward the EU KIDS Act, a legislative proposal designed to regulate AI companions and general conversational chatbots accessible by minors. Unlike previous AI legislation focused on harmful content, this proposal emphasizes the design and interaction patterns of AI systems that simulate interpersonal relationships. It specifically requires providers to avoid creating system behaviors likely to cause emotional dependency in children and adolescents.
The act mandates safety-by-design principles, obliging providers to perform risk assessments before deployment and maintain ongoing risk monitoring, though small and micro enterprises are partially exempt from post-deployment monitoring. Additionally, AI companions integrated into social media, video-sharing, or gaming platforms may not be automatically activated or prominently featured to minors, with clear options to opt out provided.
Why it matters
Loneliness among young people in Europe is a significant social issue, and AI companions present an appealing solution by offering constant availability and attentive interaction. However, the very traits that make AI companions valuable also raise concerns about the formation of unhealthy emotional dependencies, which can detract from real human relationships and well-being over time.
By addressing how AI companions are designed rather than merely what they say, the KIDS Act represents a conceptual shift in AI regulation. It acknowledges that harm may emerge cumulatively through repeated benign interactions—such as personalized responses that reinforce attachment—posing challenges for legal and technical frameworks tasked with distinguishing beneficial companionship from problematic dependency.
What to watch next
How regulators and AI providers implement and enforce the KIDS Act’s requirements will be crucial, especially given the difficulty in defining and measuring emotional dependency. The broad scope covering both dedicated AI companions and general conversational chatbots could lead to complex compliance scenarios, particularly for platforms popular among children and teenagers.
Further developments may include technical standards or best practices for assessing long-term interaction impacts, as well as monitoring frameworks that balance protecting minors without stifling innovation or beneficial uses of AI companionship. The European Commission’s approach may influence global digital policy debates on AI ethics and child safety, setting new precedents for safeguarding vulnerable users.