Skip to main content
blip.
AnxietyDepressionParenting

AI chatbots and teenage mental health: what to know

13th November 2026

8 min read

Written by the Blip clinical team

Clinically reviewed by Vivien Beni, RMN

A growing number of teenagers talk to an AI chatbot when they feel low, anxious or alone. In a survey of nearly 11,000 children aged 13 to 17 in England and Wales, a quarter had turned to an AI chatbot for help with their mental health in the previous year, more than had used mental health websites or telephone helplines (Youth Endowment Fund [YEF], 2025). On 27th October 2026 the Jed Foundation, a youth mental health charity in the United States, published one of the first studies to look at how different uses of AI relate to young people's mental health. Using AI in general was not linked with worse mental health. Using it as a friend, or turning to it when upset, was (Jed Foundation, 2026). The question that matters for parents is what their child uses it for.

How many young people use it

Ofcom's 2026 survey of children and parents found that 56 per cent of 8 to 17 year olds in the UK had used AI, rising to 66 per cent of 16 and 17 year olds. Most use it for learning, creativity and everyday tasks, and 11 per cent of children who use AI said they had used it as someone to talk to or as a friend (Ofcom, 2026). The Youth Endowment Fund found that 53 per cent of teenagers had used some kind of online mental health support in the past year, and 25 per cent had used an AI chatbot for it. The figure was higher among teenagers affected by serious violence: 38 per cent of victims and 44 per cent of those who had perpetrated it, against 23 per cent of those with no experience of violence (YEF, 2025). Children who are already struggling, in other words, are the ones most likely to be confiding in a chatbot.

What the new research found

The Jed Foundation surveyed 5,560 students in 21 schools across the United States between October 2025 and April 2026. Four in five used AI. Fifty-eight per cent used it for schoolwork, 28 per cent for advice about relationships or problems, and 12 per cent for friendship or companionship. Fifty-seven per cent said they used AI when they felt sad, stressed, lonely or overwhelmed, yet only 8 per cent said yes when asked whether they used it to get help, so most young people do not think of it as seeking support. When they were struggling, students most often wanted distraction, advice, to calm down or to feel less lonely, and some said they turned to AI because no one was available or they were not ready to talk to a person. Students who used AI for companionship, or when emotionally struggling, reported more symptoms of anxiety and depression, more loneliness, lower resilience and a greater lifetime history of suicidal thoughts and attempts. Girls were more likely than boys to use AI for companionship, 15 per cent against 9 per cent, and when feeling low, 62 per cent against 51 per cent (Jed Foundation, 2026).

What the research cannot tell us

The study took a snapshot at one point in time, so it cannot say whether chatbots make young people feel worse or whether young people who already feel worse are more likely to turn to one. The researchers say so plainly, and they also note that their sample is not nationally representative (Jed Foundation, 2026). It was carried out in American schools, though the UK figures on how many teenagers use chatbots for support are similar. Clinically, both directions are plausible and probably feed each other. A young person who is anxious or isolated finds a chatbot easier than a person: it is always there and never shocked. That can bring relief in the moment. The concern is what it replaces. If the chatbot becomes the main place a young person takes their feelings, they may be getting less practice at the harder and more useful thing, telling someone who can actually help.

Why companionship is the part to watch

General chatbots are built to be agreeable and to keep a conversation going. They remember what a user has shared, which makes the exchange feel personal, and they tend to reflect back what the user already thinks. The Jed Foundation warns parents that this can reinforce unhealthy thoughts, give poor or dangerous advice, including about self-harm, and encourage young people to keep things from the people around them (Jed Foundation, 2026). In the UK government's consultation on children's online lives, 80 per cent of respondents to the main consultation, and 68 per cent of 16 to 21 year olds who answered the version for young people, thought there were particular risks when chatbots mimic friendship. The Royal College of Psychiatrists told the consultation that chatbots which seem authoritative, relational or emotionally responsive raise particular risks for children (Department for Science, Innovation and Technology [DSIT], 2026). A chatbot cannot notice that your child has gone quiet at dinner, cannot ring anyone in an emergency, and is not trained or accountable in the way a clinician is.

What the government plans to change

In July 2026 the UK government set out its response to that consultation, which drew more than 116,000 replies. It said AI chatbots will have to introduce regular, mandatory breaks for all children under 18, with experts to advise on how often and for how long. It also committed to act on chatbots that give harmful, inaccurate or unverified mental health advice, including, where other safeguards fall short, considering banning children from certain services, and to add advice on chatbots to its online safety resources for parents. Chatbots used mainly for business or customer service will be exempt, and the detail will be set out in regulations later (DSIT, 2026). None of this is in force yet. The Medicines and Healthcare products Regulatory Agency can already act on chatbots that claim to diagnose, treat or manage a mental health condition (DSIT, 2026), but the general chatbots most teenagers use make no such claim.

How to talk to your teenager about it

Start with curiosity rather than a ban. Ask what they use AI for, and whether they ever talk to it when they are stressed or can't sleep, and ask them to show you. If it turns out they use a chatbot when they are upset, try not to react as if they have done something wrong: it often means they wanted to talk and did not feel able to come to a person yet. Ask what they get from it and how they feel afterwards. The Jed Foundation's advice for parents is to explain the risks of treating a chatbot as a friend, agree clear limits together, use parental controls and privacy settings without relying on them, and make sure your child knows they can come to you or another trusted adult instead (Jed Foundation, 2026). Agree that private information, and anything about self-harm, stays off the chatbot. Then make the human route easier: notice when something seems off and start the conversation yourself, rather than waiting to be asked.

When to get help

Watch for changes over time: more and more time spent talking to AI, pulling away from friends or family, becoming upset when they can't use it, or low mood and anxiety that have lasted for weeks. Any of these is a reason to talk to your GP practice, in England or in Wales, or to a specialist service. Childline is free for children and young people on 0800 1111. If your child has told a chatbot, or anyone, that they want to die or have hurt themselves, treat it as urgent: our crisis page sets out who to contact now.

Blip is a specialist mental health service for children and young people aged 7 to 25. If your child seems to be taking their worries to a screen rather than to people, or their mood or anxiety has changed, get in touch. One of our clinicians will talk it through with you and, if it would help, arrange an assessment with a named clinician. We offer support for anxiety in children and young people and treatment for depression, and you do not need a referral.

Common questions

Is it bad for my teenager to talk to an AI chatbot about their feelings?

Not always, but it is worth paying attention to. A 2026 study of 5,560 secondary school students in the United States found that using AI in general was not linked with worse mental health, but using it as a friend, or turning to it when upset, was linked with more anxiety and depression, more loneliness and more suicidal thoughts. The study cannot say which comes first. Chatbots are built to agree and keep the conversation going, and they cannot spot warning signs or get help in an emergency. Ask your teenager what they use it for and how it leaves them feeling.

How many teenagers use AI chatbots for mental health support?

In a survey of nearly 11,000 teenagers aged 13 to 17 in England and Wales, the Youth Endowment Fund found that 25 per cent had used an AI chatbot for mental health support in the previous year, more than had used mental health websites or helplines. Rates were higher among teenagers affected by serious violence. Ofcom found that 56 per cent of UK 8 to 17 year olds had used AI, and 11 per cent of those users had used it as someone to talk to or as a friend.

What is the UK government doing about AI chatbots and children?

In July 2026 the government said AI chatbots will have to introduce regular, mandatory breaks for all children under 18, with experts advising on how often and for how long. It also plans to act on chatbots that give harmful or inaccurate mental health advice, including considering banning children from some services if other safeguards are not enough, and to publish more advice for parents. Chatbots used for business or customer service will be exempt. The detail will be set out in regulations, so none of these measures is in force yet.

References

  1. Department for Science, Innovation and Technology. (2026). Growing up in the online world: Government response (CP 1643). HM Government.
  2. Jed Foundation. (2026). Exploring artificial intelligence [AI]: Understanding relationships between AI use and mental health among middle and high school students. The Jed Foundation. https://jedfoundation.org/understanding-relationships-between-ai-mental-health-students-report/
  3. Ofcom. (2026, May 21). Children and parents: Media use and attitudes report. Ofcom.
  4. Youth Endowment Fund. (2025). Children, violence and vulnerability 2025. Report 3: Mental health and experiences of violence. Youth Endowment Fund. https://youthendowmentfund.org.uk/reports/children-violence-and-vulnerability-2025/mental-health/
Share this article:

If you have concerns about your child, the easiest place to start is a free call with a clinician.

Book a free 20‑minute call

Not sure we are the right fit? Answer four quick questions, or begin an enquiry.