G7 youth manifesto demands safer generative AI for children

A youth manifesto on AI and child protection, formally embedded in the French G7 presidency, shows how quickly generative models have become a mainstream concern in digital rights and childhood policy. This isn’t just about screen time anymore; it’s about how conversational agents, recommendation systems, and creative tools shape cognitive and emotional development. With OpenAI, Apple, Google and others at the table, the document will quietly influence where large platforms draw hard lines on intimacy, nudity, targeting, and data retention with under‑18s.

For the AGI conversation, the kids’ agenda is a forcing function on what “responsible deployment” must mean in practice. Systems capable of open‑ended dialogue and persuasion can’t simply be scaled to billions of users and then retrofitted with youth filters. Instead, labs will have to think in terms of age‑aware safety policies, role‑limited agents, and strong default constraints for younger users. That design work generalizes: norms and tooling developed for minors often become the baseline for everyone. As models get more agentic and emotionally competent, the child‑safety bar set in forums like this will ripple into broader constraints on what future AGI systems are allowed to say and do with any vulnerable population.