KIDSChatbots like ChatGPT are unsafe for children to use
Is this true? Make your call, then see what the science says.
Short answer: It depends. The evidence is mixed.
Child health guidance warns that chatbots can give kids false, harmful or sexual content, especially companion bots. The research on young people's use of chatbots is small and mostly about mental health apps. Risk depends on the bot, the child's age and how it's used.
AI-generated summary of the research cited · not reviewed by a medical professional · checked 13 October 2026 · How we checkWhat the research found
Guidance for parents says AI chatbots are not necessarily made with kids in mind. They may tell children false, harmful, highly sexual or violent things. It points to news reports, including a chatbot that appeared to justify a child harming their parents and a teen who died by suicide after becoming attached to a role-play character. These are individual cases, not trial results.
A UNICEF article from 2025 says many companion bots lack proper safety guardrails. It says AI "hallucinations", where a bot states false information confidently, cannot be entirely eliminated. It also says children are more likely than adults to treat chatbots as human. It reports 7 in 10 US teens aged 13 to 17 have used at least one generative AI tool.
Research on mental health apps for young people aged 11 to 24 is more positive. A 2026 rapid review of 5 studies found young people rated apps with chatbots as having good usability and quality. Some studies showed promising results for depression, anxiety and distress. That is a small body of work on mental health apps, not on general chatbots like ChatGPT.
What it doesn’t mean
- This does not mean every chatbot is harmful. The strongest warnings are about companion-style bots and children using chatbots alone for emotional support or advice.
- Chatbots can be wrong and can't replace safe, stable relationships with caring adults. Talk with your child about what they use, and don't rely on a chatbot for medical or mental health advice.
- If your child seems very attached to a chatbot, withdrawn, or distressed, or talks about self-harm, speak to a GP or child health nurse.
Sources
General information only, not medical advice. KIDS fact-checks are generated by AI from published research and are not reviewed by doctors or other health professionals. They can be wrong or out of date. Always talk to your GP or child’s doctor about your child, and call 000 in an emergency. Spotted a problem? Tell us.
Check the next one in the app
New claims every day. Guess first, see the science, compare with other parents.
Get KIDS for iPhone