Children Are Adopting AI Three Times Faster Than Adults

UNICEF's finding describes a cohort moving ahead of every safeguard designed to protect it, and ahead of the adults meant to supervise its use.

Sign up for my Substack daily AI newsletter here.

See my AI Training course portfolio for corporate Business Leaders here.

Article link: https://johanosteyn.substack.com/p/children-are-adopting-ai-three-times?r=73gqa&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/

Most of the governance conversation this year has focused on boards, regulators, and enterprise risk registers. Meanwhile, the group adopting these tools fastest has no board, no regulator and no seat anywhere. They are children, and the evidence about what this is doing to them is only beginning to arrive.

CONTEXT AND BACKGROUND

UNICEF published a statement ahead of the first Global Dialogue on AI Governance, drawing on surveys of roughly 1,000 internet-using children aged 12 to 17 and 1,000 of their parents or caregivers in each of ten countries: Armenia, Brazil, Colombia, Dominican Republic, Jordan, Mexico, Montenegro, North Macedonia, Pakistan and Serbia. It estimates that at least 20 million children in those countries have used AI, with many outpacing adults by adopting it at rates more than three times faster. An estimated 13 million use it to support learning and homework. More than 2 million, or one in ten, say they turn to it for advice on things that worry them. UNICEF states plainly that as children’s use accelerates, the rules governing it, including protections for children, are struggling to keep up, and that children are more exposed to these systems while having far less power to avoid or challenge them. Its assessment of the present moment is that a generation is growing up inside a global experiment. The finding returned to circulation this week when more than 600 young people gathered at UN Headquarters on International Youth Day and presented a youth-led declaration on AI policy and governance.

INSIGHT AND ANALYSIS

Two things in that data deserve separating. Thirteen million children using AI for homework is an education problem, and it is the one schools and parents are arguing about. Two million turning to it for advice on things that worry them is something else entirely, and it is not what these systems were built for. A tool optimised to be helpful and agreeable is a poor confidant for a frightened child, and the evidence about cognitive development, emotional dependency and exposure to harm is, in UNICEF’s own words, just emerging. The children themselves are not naive about this. Across the ten countries, a third reported concerns about AI being used to scam and trick people or spread misinformation, and a quarter feared having their images or videos manipulated into sexually explicit deepfakes. That last figure should end any argument that this is a moral panic imported from adults. UNICEF’s conclusion is that too many systems are reaching children with no guardrails, with safety seemingly an afterthought.

IMPLICATIONS

South Africa was not among the countries surveyed, and that absence is itself informative. We have no comparable national data, no finalised AI policy following the withdrawal of the April draft, and a working-age adoption rate of 23.1 percent that tells us nothing about anyone under eighteen. If the pattern holds here, and there is no obvious reason it would not, then South African children are using these systems at rates their parents, teachers and the Department of Basic Education cannot currently measure. UNICEF’s recommendations are directed at governments and companies, and they include strengthening laws and corporate accountability to stop AI-enabled sexual exploitation, requiring systems designed with maximum safety and transparency, and building AI literacy for children and their caregivers together rather than separately. None of that requires a national policy to begin.

CLOSING TAKEAWAY

Business leaders reading this have spent the year on procurement, cost and compliance, which is the right work and is not the only work. The same executives approving enterprise AI governance frameworks go home to households where the fastest adopters in the country are asking a language model about things that worry them. The governance question is not confined to the office. UNICEF calls this a decisive moment on the grounds that the choices made now will shape children’s safety, privacy and wellbeing for decades. The uncomfortable part is that the choices are being made by default, by companies optimising for engagement, in the absence of anyone with authority saying otherwise.

Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net