Children are embracing AI far faster than adults, and that speed is now raising alarms about privacy, safety, and the way young users think. UNICEF says the technology is already woven into everyday life for many children, but the safeguards around it are still lagging behind.
Based on a survey of 1,000 internet users aged 12 to 17 and 1,000 parents across 10 countries, UNICEF said at least 20 million children have already used AI. Around 13 million of them use it to help with schoolwork and homework.
AI Is Becoming Part of Childhood
The most worrying detail is that more than 2 million children, or 1 in 10, said they use AI to look for solutions to personal problems. UNICEF said this shows AI has already become part of childhood in multiple countries, even though protection measures have not kept pace.
The organization said many AI systems reach children without clear safety barriers. In its report, UNICEF also warned that safety is still too often considered only after the technology has already spread widely.
| UNICEF Finding | Figure | What It Means | Context |
|---|---|---|---|
| Children surveyed | 1,000 internet users aged 12-17 | Views from young users were directly collected | Across 10 countries |
| Parents surveyed | 1,000 parents | Parents were included in the same survey | Across 10 countries |
| Children who have used AI | At least 20 million | AI use among children is already widespread | UNICEF estimate from the survey |
| Children using AI for schoolwork | Around 13 million | Study support is a major use case | Homework and learning help |
| Children using AI for personal problems | More than 2 million, or 1 in 10 | AI is also being used for sensitive personal issues | Raises safety concerns |
Deepfakes, Hoaxes, and Cognitive Decline
Children themselves appear aware of some of the dangers. One-third said they worry AI could be used to trick people or spread hoaxes, while a quarter fear their photos or videos could be turned into vulgar sexual deepfakes.
Those fears overlap with findings from researchers. A 2025 study by MIT Media Lab warned that heavy dependence on AI may trigger “cognitive atrophy,” meaning a decline in critical thinking skills.
Another study from Italian researchers, The brain side of human-AI interactions in the long-term, published in January 2026, pointed to a similar risk. It said AI can encourage cognitive delegation that weakens information processing, problem-solving ability, and even a child’s moral compass.
New Pressure on Rules for Harmful Content
Concerns are also growing over the misuse of AI to generate nude images. The latest case involved the generative feature on Elon Musk’s Grok AI, which prompted Britain’s communications regulator, Ofcom, to step in.
Grok later restricted that feature to paying subscribers only, but a spokesperson for British Prime Minister Keir Starmer criticized the move because it only turned an illegal image-making tool into a premium service.
If access like this is available to children, age-based porn-blocking rules in many countries could become ineffective. UNICEF is therefore urging governments and private companies to include children’s rights in global AI governance.
The organization wants stronger criminal laws against AI-based sexual exploitation, better digital literacy, and a requirement for companies to design systems that are transparent and safe. UNICEF said the choices made now will shape children’s safety, privacy, and well-being for years to come.
As AI becomes more familiar to young users, the central issue is no longer whether children will encounter it, but whether the systems they use will be built with proper protection from the start.
