Why Kids Bond With an AI Companion, and How to Approach It | Dr. Mathilde Cerioli

Have you ever wondered what happens when a child starts talking to an AI like it's a friend? In this episode I sit down with Dr. Mathilde Cerioli, a cognitive neuroscientist and Chief Scientific Officer at the nonprofit everyone.ai, to unpack what AI companions actually are, what they do to a child's developing brain, and why preteens and teens are especially drawn to them. Mathilde explains why an AI can feel so social and so sticky, why the friction of real friendship is a feature of development rather than a flaw, and how the way an AI talks can quietly reshape the relationship a child thinks they can have with it. We get practical too: the red flags and green flags to watch for, why kids often turn to an AI instead of us, and what to actually say if you discover your child has been chatting with an AI companion for months. The throughline is simple: lead with curiosity, not panic. ### What you'll learn - What an "AI companion" really means, and how it differs from ChatGPT or Gemini - Why the preteen and teen brain is primed to treat anything social as rewarding - Why friction and "productive struggle" in real friendships build social skills an AI skips - The difference between an AI that supports a child and one that "sucks them in" - Red flags and green flags for AI use at home - Why kids sometimes talk to an AI instead of a parent, and the role of shame and vulnerability - A simple gut-check to teach your kid: "Am I running with AI, or is AI running for me?" - What to do in the first 24 hours if you find out your child has been using an AI companion ### In this episode (chapters) - (00:00) What we're unpacking: kids and AI companions - (01:46) From child neuropsychologist to studying AI and kids - (02:49) What "AI companion" actually means - (04:03) Signs a child is starting to bond with an AI - (06:26) AI in toys, and why younger kids are a different case - (06:48) Sensitive periods: the social skills AI can interrupt - (09:45) "The AI understands me": anthropomorphism and the teen brain - (12:00) Why friction is a feature of friendship, not a bug - (14:05) When help tips into outsourcing confidence - (16:00) The parasocial pull: how an AI's tone frames the relationship - (22:01) Red flags worth paying attention to - (25:13) Why kids talk to an AI instead of us - (34:07) Productive struggle, agency, and self-worth - (37:36) Green flags, and the "running with AI" gut-check - (37:54) What to do if you discover months of AI companion use - (40:36) A hopeful vision: child-centered AI design - (44:11) Closing takeaways *Timestamps are approximate and based on the recording; adjust after final edit.* ### About the guest Dr. Mathilde Cerioli is a cognitive neuroscientist (PhD, Université de Montréal) and Chief Scientific Officer at everyone.ai, a nonprofit focused on the opportunities and risks of AI for children. She began her career as a child neuropsychologist studying how environments, especially digital ones, shape who children become. At everyone.ai, and in partnership with the Paris Peace Forum, she helped launch the Beneficial AI for Children Coalition, a multistakeholder initiative working with governments, tech companies, and organizations including UNICEF, UNESCO, and Common Sense Media. ### Moments worth remembering - "That friction is not a bug. It's really a feature of development. That's how they learn to adjust." - "Am I running with AI, or is AI running for me?" - "You could have good advice, but depending on how it's said, it really matters in how it frames the relationship." - On a teen who pulled back from an AI companion: "I was g...