Blab AI
UNICEF Finds 20 Million Children Using AI, Highlights Governance Gap
On 30 June, UNICEF released a striking new analysis that shows at least 20 million children aged 12‑17 across ten countries have already turned to artificial‑intelligence (AI) tools. The study, part of the Disrupting Harm Phase 2 project, reveals that kids are adopting AI more than three times faster than adults in the same age brackets.
Researchers surveyed roughly 1,000 internet‑using children and 1,000 parents or caregivers in each of Armenia, Brazil, Colombia, the Dominican Republic, Jordan, Mexico, Montenegro, North Macedonia, Pakistan and Serbia. UNICEF and Ipsos then weighted the national figures against United Nations population data to arrive at the global estimate.
Key findings paint a vivid picture of how AI is weaving itself into young lives. About 2 million children—roughly one in ten—use AI to seek advice on personal concerns, while 13 million turn to the technology for help with schoolwork and homework. Yet a third of the children surveyed voiced worries that AI could be used to scam people or spread misinformation, and a quarter feared that their own images or videos could be manipulated into sexually explicit deepfakes.
UNICEF’s statement underscores a stark reality: children are exposed to AI systems, their underlying business models and the ways their data is harvested, yet they have little power to avoid or challenge those systems. The agency notes that weak governance is felt most acutely by children, who will live with the consequences for longer than adults.
The report arrives ahead of the first Global Dialogue on AI Governance and follows a wave of high‑profile debates over child safety online. In the United States, Congress is considering legislation to protect children from online harms, while the state of Florida has filed a lawsuit against OpenAI alleging that the company’s chatbot poses safety risks for young users.
UNICEF’s recommendations are not new, but the scale of the adoption figure underscores the urgency. The agency calls for: 1. More research into AI’s effects on child development; 2. Stronger laws against AI‑enabled sexual exploitation; 3. Safety and transparency built into AI systems by design; 4. Wider AI literacy support for children and caregivers; 5. Investment in connectivity to prevent widening gaps between countries.
The report distinguishes between adoption speed and volume. While children are adopting AI tools rapidly, many adults are still learning to use generative AI effectively. This pattern mirrors workplace adoption data tracked by industry analysts.
Policy responses such as Malta’s national AI literacy course illustrate one approach: pairing access to AI tools with structured teaching before children and parents are left to navigate the technology on their own.
UNICEF has not set a timeline for implementing its recommendations, nor identified which governments or companies are lagging the most. The agency’s statement concludes that the window for shaping rules is closing at the same pace that children are opening AI apps.
The findings highlight a growing mismatch between the pace of AI adoption among young people and the development of safeguards that protect them. As AI tools become more integrated into education, social media, and everyday life, the need for targeted governance that accounts for children’s unique vulnerabilities is increasingly clear.
The report is part of the broader Disrupting Harm research program, funded by Safe Online and conducted in partnership with ECPAT International and INTERPOL. The full country reports are available on UNICEF’s Innocenti website.
In short, 20 million children are already using AI, and the regulatory and protective frameworks that should accompany that use are lagging behind. The next steps will involve translating UNICEF’s five‑point call into concrete policy, industry, and educational actions that keep pace with the technology’s rapid spread among young users.
Researchers surveyed roughly 1,000 internet‑using children and 1,000 parents or caregivers in each of Armenia, Brazil, Colombia, the Dominican Republic, Jordan, Mexico, Montenegro, North Macedonia, Pakistan and Serbia. UNICEF and Ipsos then weighted the national figures against United Nations population data to arrive at the global estimate.
Key findings paint a vivid picture of how AI is weaving itself into young lives. About 2 million children—roughly one in ten—use AI to seek advice on personal concerns, while 13 million turn to the technology for help with schoolwork and homework. Yet a third of the children surveyed voiced worries that AI could be used to scam people or spread misinformation, and a quarter feared that their own images or videos could be manipulated into sexually explicit deepfakes.
UNICEF’s statement underscores a stark reality: children are exposed to AI systems, their underlying business models and the ways their data is harvested, yet they have little power to avoid or challenge those systems. The agency notes that weak governance is felt most acutely by children, who will live with the consequences for longer than adults.
The report arrives ahead of the first Global Dialogue on AI Governance and follows a wave of high‑profile debates over child safety online. In the United States, Congress is considering legislation to protect children from online harms, while the state of Florida has filed a lawsuit against OpenAI alleging that the company’s chatbot poses safety risks for young users.
UNICEF’s recommendations are not new, but the scale of the adoption figure underscores the urgency. The agency calls for: 1. More research into AI’s effects on child development; 2. Stronger laws against AI‑enabled sexual exploitation; 3. Safety and transparency built into AI systems by design; 4. Wider AI literacy support for children and caregivers; 5. Investment in connectivity to prevent widening gaps between countries.
The report distinguishes between adoption speed and volume. While children are adopting AI tools rapidly, many adults are still learning to use generative AI effectively. This pattern mirrors workplace adoption data tracked by industry analysts.
Policy responses such as Malta’s national AI literacy course illustrate one approach: pairing access to AI tools with structured teaching before children and parents are left to navigate the technology on their own.
UNICEF has not set a timeline for implementing its recommendations, nor identified which governments or companies are lagging the most. The agency’s statement concludes that the window for shaping rules is closing at the same pace that children are opening AI apps.
The findings highlight a growing mismatch between the pace of AI adoption among young people and the development of safeguards that protect them. As AI tools become more integrated into education, social media, and everyday life, the need for targeted governance that accounts for children’s unique vulnerabilities is increasingly clear.
The report is part of the broader Disrupting Harm research program, funded by Safe Online and conducted in partnership with ECPAT International and INTERPOL. The full country reports are available on UNICEF’s Innocenti website.
In short, 20 million children are already using AI, and the regulatory and protective frameworks that should accompany that use are lagging behind. The next steps will involve translating UNICEF’s five‑point call into concrete policy, industry, and educational actions that keep pace with the technology’s rapid spread among young users.
