The European Union is set to introduce strict regulations on chatbots, aiming to prevent children from forming emotional ties with AI companions. The EU Kids Act, which also targets social media and video-sharing platforms, mandates that children under 13 must be supervised when engaging with chatbots. Digital platforms will face hefty fines if they fail to implement child-safety measures.

Under the new rules, chatbots will not be allowed to retain memories of interactions, as the EU fears this could lead to the accumulation of sensitive data on minors. The legislation also prohibits platforms from automatically activating or encouraging the use of AI chatbots on their services, particularly for minors.

European Digital Rights policy adviser Simeon de Brouwer welcomes the chatbot measures but questions the age-specificity. He argues: ‘Don’t we all deserve protection from addictive and emotionally manipulating AI?’ The EU’s Digital Fairness Act, yet to be unveiled, is expected to broaden the scope of restrictions on addictive design features.

President Ursula von der Leyen highlighted the dangers of chatbots, noting that around half of Europeans aged 11 to 25 have discussed personal matters with AI chatbots. The Commission is wary of chatbots’ potential to give minors harmful advice, particularly on mental health and personal development. Concerns are also growing about the cognitive impacts and deskilling effects of AI chatbots on children's educational performance.