Children Adopting AI Three Times Faster Than Adults: UNICEF

An estimated 20 million children across ten countries have already used artificial intelligence, adopting the technology more than three times faster than adults. That stark finding comes from a UNICEF statement released on June 30, timed to coincide with the first Global Dialogue on AI Governance. The message is clear: the rules meant to shield young users online are lagging far behind their rapid adoption.

The data emerges from Disrupting Harm Phase 2, a research initiative led by UNICEF’s Office of Strategy and Evidence at Innocenti, in partnership with ECPAT International and INTERPOL, with funding from Safe Online. Fieldwork spanned ten nations including Armenia, Brazil, Colombia, the Dominican Republic, Jordan, Mexico, Montenegro, North Macedonia, Pakistan, and Serbia. In each country, roughly 1,000 internet-using children aged 12 to 17 and 1,000 of their parents or caregivers were surveyed. UNICEF and Ipsos then weighted the national figures against UN population data to produce global estimates.

Two key findings anchor the release. More than two million children, roughly one in ten, said they turn to AI for advice on worries or concerns. A separate estimate places 13 million children using AI tools for schoolwork and homework, a more mundane but far larger use case. Yet UNICEF’s framing is not celebratory. “Children are more exposed to AI systems, including how they are designed, their underlying business models, and how their own data is used, yet have far less power to avoid or challenge them,” the organization stated.

The agency argues that children feel the effects of weak governance first and live with the consequences longest, while most AI governance in practice does not treat them as a distinct group. This tension between fast uptake and thin protection has already shaped other battles over kids’ online safety legislation in the US Congress and Florida’s lawsuit against OpenAI over chatbot safety for young users.

Children are not naive about the risks, according to the data. A third said they worried about AI being used to scam people or spread misinformation. A quarter feared having their own images or videos manipulated into sexually explicit deepfakes, a concern UNICEF has raised before in a separate statement on deepfake abuse from the same research program.

UNICEF’s ask to governments and the private sector is a five-point list: more research into AI’s effects on child development, tougher laws against AI-enabled sexual exploitation, safety and transparency built into AI systems by design, wider AI literacy support for children and caregivers, and investment in connectivity so the gap between countries does not widen further. None of this is new territory for the agency, though the scale of the adoption number underneath the request is.

The three-times faster adoption figure describes speed rather than volume of use, a distinction UNICEF’s wording does not fully spell out. Many adults are still finding their footing with generative tools, a pattern TNW has tracked in workplace adoption data. Programs like Malta’s national AI literacy course suggest one policy direction, pairing access with structured teaching before children and parents are left to work it out alone.

UNICEF has not put a date on when any of its recommendations might be adopted, nor named which governments or companies it considers furthest behind. What it does say plainly is that the window for shaping the rules is closing at the same speed children are opening the apps.