A boy caught a praying mantis in the yard and watched it turn brown. When he asked ChatGPT why, the AI confidently told him the insect was pregnant. A few days later, Manty died. The child was devastated, but his mother, the technology journalist Joanna Stern, saw a different lesson crystallise: children need to question everything a machine tells them. That small domestic drama, recounted in an essay on her year of living with AI, captures a tension now rippling through parliaments and living rooms across several continents. The tools meant to inform and connect the young are also, in ways no one fully anticipated, reshaping the raw material of childhood itself.

Viewed from Washington, the legislative machinery is accelerating. The US House of Representatives has just passed the Kids Internet and Digital Safety Act with broad bipartisan support, requiring platforms to offer minors ways to limit addictive features and to adopt policies against sexual exploitation. The Senate, for its part, has been sitting on a more stringent bill since 2024 that would impose a “duty of care” on social media companies toward young users. In Brasília, a new legal framework known as the Digital ECA has already determined that tech products likely to be accessed by children must adopt default settings that prevent compulsive use. Australia, meanwhile, is doubling fines for tech giants that refuse to hand over internal documents proving they are enforcing a social media ban for under-16s; the eSafety Commissioner remains “concerned” about Facebook, Instagram, Snapchat, TikTok and YouTube. And in the United Arab Emirates, a prohibition on social media use for children under 15 is being paired with a new regulatory framework and awareness programmes, while 15,000 licensed content creators from more than 90 nationalities now operate under a permit system designed, officials say, to raise standards rather than restrict creativity.

Yet the push to verify age is creating an unexpected mirror effect. In the United Kingdom, where platforms must be able to tell who is under 16, the practical consequence is that adults may have to prove they are not children. Facial age estimation, credit card checks, and repeated verification across multiple services could become routine, building vast databases of biometric and financial information. Analysts in London note that a policy aimed at protecting the young is quietly redrawing the boundary of adult privacy, forcing millions to hand over more personal data than ever before simply to access services that once required only a username and password.

Beneath the regulatory architecture lies a deeper cultural anxiety about what is being lost. Educators and parents in the Gulf and beyond report a rise in referrals for speech and language delays, social communication challenges, and emotional regulation concerns. They describe groups of friends who spend more time creating content for social media than talking to one another, and a generation for whom social approval is no longer a matter of belonging but a set of metrics—likes, comments, followers—that can be checked at any hour. Stern’s experiment with an AI-powered stuffed animal left her relieved when her son showed no interest; a robot dog, by contrast, made him cry when it had to be returned. “No AI friends or pets,” she concluded, a line that reads less like a tech reviewer’s verdict than a parent’s quiet act of resistance.

In the end, the most telling image may be the one Stern offers of her sons after dinner. They are not clamouring for screens or chatbots. They are playing a game they call “robots,” pretending to be the very machines that so many adults are now trying to fence off. It is a reminder that children have always metabolised the world around them through make-believe, and that the real experiment is not in the technology but in the human relationships that survive alongside it.