How Many Kids Use AI? What the Latest Surveys Say

About two in three. In a July 2026 Thorn survey of US 9 to 17 year olds, 67% had used an AI chatbot, including 60% of 9 to 12 year olds. Ofcom's UK figure is lower, 56% of 8 to 17s, partly because surveys word the question differently. Use is common, but a number cannot tell you whether your child's use is healthy, and your school's rules come first.

What do the surveys actually say?

Four recent surveys give the clearest picture, two American and two British. They do not measure quite the same thing, so I have kept them separate rather than blending them into one headline.

Read as a group, these say that somewhere between a half and three-quarters of school-age children have used an AI chatbot, that this is true of most children in the 9 to 12 range as well as teenagers, and that use rises with age.

Why do the numbers differ?

Because the questions differ. Thorn asked about "a computer program that you can text or talk with whenever you want, like ChatGPT or Character AI." Pew asked about AI chatbots. Ofcom's page counts children who use "AI", which may be a wider net than "chatbot" because its page does not define the term. Thorn's panel is an online opt-in sample weighted to match US census figures, and Thorn's own report describes it as a nonprobability sample, so its percentages are good estimates, not exact counts.

Every figure above also counts a child who has used a chatbot at least once. None of them says how often, for how long, or on whose account.

My interpretation: the plain summary is "most, and the exact share depends on how you ask." If a headline gives you one precise number for all children, treat it with some suspicion.

What about children under 13?

Most of the major chatbots set a minimum age. OpenAI's help center says "ChatGPT is not meant for children under 13" and that it requires "children ages 13 to 18 obtain parental consent before using ChatGPT." Our guide to the age limits on the main AI tools sets out how each company handles it.

The surveys suggest those limits are not stopping many children. In Thorn's survey, 43% of 9 to 12 year olds said they had used some app, website or platform before they were old enough, and across all ages the most common route, at 63%, was a fake birthday. That question covers platforms in general, not chatbots specifically, so please read it as a signal about age limits, not a chatbot count.

The surveys here start at age 8 or 9, so they say nothing reliable about younger children. For that age, our guide to AI and young children is a better starting point than a statistic.

Which chatbots are they using?

Among the Thorn respondents who had used a chatbot, 78% had used ChatGPT, 41% Gemini, 28% Meta AI, 16% Snapchat's My AI, 15% Character AI, 8% Grok and 3% Claude. Children could pick more than one, so the figures add up to more than 100%.

The practical point is that the tool most children have used is a general-purpose one, built for adults, and not a children's product. If your child has used "an AI", it was most likely ChatGPT.

What are they using them for?

Schoolwork and finding things out come first. Among UK children who had used a chatbot, Internet Matters found 42% used one for schoolwork or learning, and that this rose with age. Pew found 54% of US teens had used chatbots for schoolwork and 57% to search for information. Our guide to what teenagers use AI for goes through the Pew detail.

Some uses are more personal. In Thorn's data, among children who had used a chatbot, 31% had used one to talk about or get advice on romantic feelings or relationships, 31% for help with something scary or confusing that happened to them, and 26% for information about puberty or sexual health. Thorn also reported that 39% had not used one in any of those ways. Ofcom found 11% of 8 to 17 year olds had used AI "as someone to talk to or 'as a friend'."

These are children reporting what they have done, and several of the uses are the ones that a parent would rather the child brought to a person. Nothing in these surveys shows that using a chatbot caused any problem for these children. They describe use, not harm.

What I See in Schools

The pattern I see most often in secondary schools is a gap between two pictures. The school's picture of AI comes from what happens in lessons and in the policy. The survey picture is mostly what happens at home, on a personal phone, on an account the school has never seen.

When I train teachers, the question that keeps coming up is not "do students use AI?" because everyone assumes they do. It is "what are they asking it that they would not ask us?" A statistic cannot answer that, and neither can a school. Only the child can.

A survey tells you what most children do. It cannot tell you what yours did last night, and only a conversation can.

What I Would Do at Home

This is my suggestion, not a finding. The goal is to learn what your child does without making them defend it.

Ask a specific question, not a general one. "Do you use AI?" gets a shrug. You could say: "What's the last thing you asked it?" It is easy to answer truthfully, and the answer is usually mundane, which lowers the temperature for both of you.

Say what you are not going to do. You could say: "I'm not checking up on you. I want to know what it's good for, because I don't know yet." Children answer differently when the question is curiosity and not an audit.

Match it to the age:

• Under 8: The surveys do not cover this age. Treat AI like any other screen you supervise, and use it together.
• 8 to 12: Most 9 to 12 year olds in the surveys have tried a chatbot, often on a general-purpose app that officially excludes them. Ask whose account they used. Our guide to AI for 9 to 12 year olds covers the options.
• 13 to 15: Use is common and often private. Agree where it is used and what it is for, and talk about what to do when an answer is wrong or strange.
• 16 to 18: The job moves from permission to judgment. Ask how they check what it tells them.

When should the number worry me?

It should not, on its own. Using a chatbot is normal. Things that deserve attention are different: a child who seems to prefer it to people, who is hiding it, or who is turning to it when they are upset. Our guide to AI and feelings and signs a child may be leaning on AI too much explain what to look for and what to say.

If you are worried about your child's wellbeing, a chatbot is not a clinician and neither is this page. Speak to your child's pediatrician or doctor (GP in the UK) or the school's safeguarding lead. In the US you can call or text 988 at any time, and in the UK Samaritans is on 116 123 and Childline on 0800 1111.

What to Ask the School

Schools may know less about home use than parents assume. Three questions help: does the school know how many students use AI outside lessons, what does the policy say about using it for homework, and how will parents hear about changes? Ten questions to ask your child's school about AI has the full list.

If your child's school has not spoken to parents about AI yet, that gap is worth naming, and it is the work I do with schools: training staff, setting a strategy and making the policy mean something in the classroom. It is worth sending them. I also write a newsletter that keeps teachers and parents up to date without the panic, and parents are very welcome.

Dan Fitzpatrick is a former secondary teacher and assistant headteacher who has trained more than 150,000 educators in over 30 countries. More about Dan.