Children using AI has become an ordinary feature of digital childhood, but that does not mean young people love it. In a new Common Sense Media census of American children aged nine to seventeen, 65 per cent had used AI in their free time. Among those who used it, however, only 11 per cent of teenagers and 15 per cent of younger children said they enjoyed it a lot. Frequency and enthusiasm were telling different stories.
The survey, published on 27 October 2026, does not show that children hate chatbots. Saying something is not highly enjoyable is not the same as disliking it. A child can use AI because it helps with a task, because friends mention it or because it is built into a service. The finding matters because parents and policymakers often judge a technology by how often children encounter it rather than what they get from it.
Children using AI are encountering it surprisingly young
Common Sense Media surveyed a nationally representative group of 1,211 young people in the United States. The research covered several kinds of entertainment and digital activity, including video, games, social media and generative AI. It found that AI had entered the everyday media environment of younger children, not merely the study habits of older teenagers preparing for examinations.
The organisation’s Youth AI Safety Institute analysis gives an unusually clear picture of the age pattern. It reports chatbot use among 36 per cent of nine-year-olds, 55 per cent of eleven-year-olds and 73 per cent of thirteen-year-olds. These are survey estimates for American young people, not verified accounts showing how many children registered on each service or how often every child uses one.
The speed of adoption raises a straightforward question: what does a nine-year-old understand an AI chatbot to be? Some services enforce age restrictions, yet children may encounter similar functionality through other products or shared devices. Knowing the name of a popular chatbot is different from understanding why its answer can be wrong, why personal details matter or why an apparently friendly response is generated by software.
Parents also face a moving target. The same broad label can describe a search-style question box, a homework helper, an image generator or a companion character that encourages extended conversation. Their purpose and risks are not identical. A useful family discussion begins by identifying what the child is actually using, rather than arguing about AI as though it were one uniform product.
The gap between use and enjoyment
The census found the widest gap between use and strong enjoyment in the AI category. A young person may return to a chatbot without treating it as a favourite pastime. That is not necessarily a failure of the technology. Many adults rely on calendars, search engines and word processors without describing them as enjoyable. The tool may be doing its job precisely because it solves a problem and then disappears.
The study’s wording matters. Only a small minority of users said they enjoyed AI a lot; it did not ask them simply to vote for or against the existence of chatbots. Some may enjoy it a little. Others may be indifferent, curious, frustrated or still experimenting. The findings should not be converted into a sweeping claim about the feelings of every young person with access to AI.
That distinction changes how adults interpret screen time. Spending an hour with a video platform often means something quite different from spending an hour using a chatbot for information, a game, schoolwork or social reassurance. The visible activity is still screen-based, but the intention behind it changes the experience. Minutes alone reveal surprisingly little about whether a particular session was beneficial or troubling.
The Institute notes that only a small share selected ChatGPT as the single platform they could not live without. Video remained much more important in that particular question. The contrast suggests that AI can become commonplace without replacing the established attractions of online entertainment. It is a reminder that every new digital feature does not automatically dominate young people’s interests simply because adults regard it as revolutionary.
A useful tool or an unwanted extra demand?
One explanation is that children use AI because a task has been placed in front of them. They may need an explanation, a draft or a quick answer. That can be helpful, but the experience may not be enjoyable in the sense of a favourite video game. A school assignment does not become entertainment merely because the student uses an advanced system to complete part of it.
Another possibility is that novelty fades quickly. The first coherent response from a chatbot can be astonishing. After repeated use, its mistakes, repetition and awkward instructions become apparent. The survey does not prove this is why the enjoyment gap exists. It does, however, challenge the assumption that growing usage is a reliable measure of satisfaction or a sign of a lasting emotional attachment.
The boundary between learning support and dependency deserves particular care. A chatbot can explain a confusing instruction, offer practice questions or give feedback on a draft. It can also produce an answer so complete that the child no longer needs to solve the problem. LiveAIWire has examined how an AI tutor can use questions to support classroom learning, an approach that treats the pupil’s understanding as the outcome rather than the polished response on the screen.
Parents do not need a technical test to notice the difference. Ask the child to explain what the chatbot helped them understand. If they can describe the underlying idea in their own words, the interaction may have been productive. If they can only paste the answer, the result is less encouraging. Such conversations will not prove an educational effect, but they may reveal whether the tool is acting as a support or a substitute.
The under-thirteen problem has not gone away
The survey’s youngest users are particularly significant because many major chatbots set a minimum age of thirteen or higher for some experiences. Reported use among younger children suggests age limits alone are not keeping the technology away from them. That does not establish that any particular provider has violated its rules, since some use may occur through other services or in supervised settings.
Families may encounter AI through devices already approved for other purposes. A familiar game, search service or school platform can introduce functions that were not present when a parent first agreed to its use. Controls that worked yesterday may not cover the new feature. The practical task is to keep checking how the product behaves, which information it requests and whether a younger user can recognise the difference between fiction and advice.
There is a wider policy debate over companion chatbots, which are designed to encourage conversation rather than simply answer a single question. LiveAIWire reported on New York’s rules for AI companion services and children. Those legal concerns are not evidence that ordinary informational use is dangerous, but they show why a parent should distinguish a homework tool from a system built to sustain an apparently personal relationship.
Appropriate boundaries depend on the child’s age and the type of service. Younger users may need shared sessions, explicit supervision or products designed for education. Older teenagers may benefit from discussions about privacy, misinformation and persuasive design. A blanket instruction to avoid all AI is unlikely to explain its role in school and wider life. Equally, unrestricted access because a tool seems useful can ignore genuine developmental differences.
The family conversation that makes more sense
Ask what the child actually does with AI, not merely whether they have tried it. Is it for homework, entertainment, asking awkward questions or creating pictures? Does it provide useful explanations, or does it frustrate them? Are there situations where they would prefer an adult, a teacher or a friend? These questions give better information than a single count of hours spent on a device.
It helps to discuss how confidently a chatbot can make mistakes. An AI answer might sound certain about a fact that is out of date or invented. Children can be encouraged to check important statements against a book, teacher or trusted website. This is a practical skill, not a reason to dismiss every answer. Learning to question software may be as important as learning how to operate it.
Privacy belongs in the same discussion. Children should know not to share passwords, private messages, precise home details or identifying information about other people merely because a chatbot asks for context. The difference between a real friend and an artificial conversational system can be explained without trying to frighten them. LiveAIWire’s report on teenage AI use and parents in France provides another example of how families are having to negotiate these new habits.
What the new census really tells us
The 2026 figures are a snapshot of American young people’s reported behaviour. They cannot tell whether AI has improved their grades, made them less social or affected their wellbeing. Nor do they establish how the same children will use the technology next year. The point of a broad media census is to show patterns, not to settle every causal question about childhood development.
Nevertheless, the enjoyment gap is a valuable corrective to the industry’s excitement. Being widely used is not proof of being loved. It may simply mean a technology has become convenient, unavoidable or embedded inside other services. Families and educators should judge it by what a child learns, whether their autonomy is protected and whether the experience is appropriate for their age.
The interesting question is no longer whether children will encounter artificial intelligence. Many already have. It is whether the adults around them can help distinguish the genuinely useful applications from those that waste time, invite oversharing or replace thinking. The survey supplies a reason to ask better questions, not a simple verdict on whether AI belongs in childhood.
About the Author
Stuart Kerr is Technology Correspondent at LiveAIWire, covering artificial intelligence, cybersecurity and the social impact of emerging technology. LiveAIWire is an independent, human-led technology publication using AI-assisted research, editorial production and original AI-assisted editorial illustrations under his direction.