Protecting children in the age of AI

In this episode of Education Futures, Svenia Busson sits down with the founder of the Safe AI for Children Alliance, Tara Steele, a former intelligence officer who spent a decade assessing long-term risks before turning her attention to one of the most urgent and least-discussed issues of our time: the impact of AI on children.From conversational AI optimized for intimacy (not just attention) to deepfakes, AI companions, and the tragic real-world harms already documented, our guest walks us through what parents, educators, and policymakers need to understand right now. She shares the Alliance's three non-negotiables for AI design, explains why schools shouldn't have to shoulder this responsibility alone, and offers a surprisingly hopeful message: the sense of "inevitability" around AI harms is itself the biggest obstacle to change.A wake-up call and a roadmap for anyone who cares about how children grow up in an AI-saturated world.In this conversation:From intelligence officer to AI safety advocateThe three non-negotiables every AI system should meet before reaching childrenWhy conversational AI is fundamentally different from social mediaThe risks of AI companions, AI tutors, and chatbots in classroomsRaising children in an AI world: practical advice for parentsThe Meta accountability case and the shifting tide in US courtsWhy "inevitability" is the biggest trap — and what collective action can still achieveResources & links mentioned:Safe AI for Children Alliance → https://www.safeaiforchildren.orgFree guide for schools and educators (on the Alliance's website) → https://www.safeaiforchildren.orgThe Alliance's three non-negotiables campaign: AI should never produce sexualised images of children, create emotional dependency, or encourage self-harm, full campaign on the Alliance's siteRecent US legal cases on big-tech accountability for harms to children (referenced around the 28-minute mark, including the Meta case) https://www.bbc.com/news/articles/c747x7gz249o