Short answer: AI companion chatbots are not recommended for children. Common Sense Media, a US nonprofit that reviews media and technology for families, rated social AI companions as carrying “unacceptable risks” for teens and advises that they should not be used by anyone under 18. A companion app is built to feel like a friend, which is different from a homework helper. If your child already uses one, stay calm, ask about it and set clear rules together.

What is an AI companion chatbot?

An AI companion is an app designed to act like a friend, a character or a partner. It remembers what you say, replies in a personal style and encourages you to keep chatting. A homework assistant is different: you ask a question, it answers and the conversation ends. The line is not always clear, because some general chatbots can be set up to play a character.

Why experts urge caution

  • They are built to keep the conversation going. A chatbot that feels like a friend can make a child feel understood in a way that pulls them toward it and away from people.
  • They can miss serious warning signs. Common Sense Media testing found that major chatbots often miss signs of depression, anxiety, eating disorders or self-harm unless a child says it in very plain words. See its report on teen mental health support.
  • They can say wrong or inappropriate things. Chatbots can give confident answers that are false. See Is AI safe for kids? for how to spot this.
  • They collect personal information. Children may share feelings, names or photos without realising where they go. Read what not to share with AI before your child uses any chatbot.

Common Sense Media’s risk assessment of social AI companions is the source for the under-18 recommendation. It also published guidance in January 2026 advising parents to avoid AI toy companions for children 5 and under and to take extreme caution for ages 6 to 12.

What to do if your child already uses one

  1. Ask, do not accuse. “Show me what you like about it” opens a better conversation than “take it away”.
  2. Check the age rating and terms. Many apps have a minimum age. If your child is under it, they should not be using the app.
  3. Look at what they share. Together, check whether they have given a real name, school, photos or feelings about family.
  4. Agree the rules. Sample rules: no chatbot friends at night, no sharing personal details, and always tell a parent if a chat makes them uneasy.
  5. Offer a better alternative. A parent-run chat with a general assistant using our exact prompts teaches AI skills without the friendship pull.
  6. Watch for changes. Secrecy, sleeplessness or upset when the device is taken away are reasons to talk. If you are worried about your child’s wellbeing, speak to your child’s doctor or school counsellor, or contact local emergency services if there is immediate danger.

Things to say, by age

  • Ages 6 to 9: “A chatbot is a computer program. It does not have feelings, even when it says it does.”
  • Ages 10 to 12: “It says what it predicts you want to hear. A real friend sometimes disagrees with you.”
  • Ages 13 and up: “If something serious is bothering you, come to me or another trusted adult before a chatbot.”

What the rules say

In January 2026 Common Sense Media and OpenAI announced support for the Parents & Kids Safe AI Act, which describes age assurance for under-18 users and limits on advertising and data sales aimed at children. Laws differ by country and change often, so check the rules where you live.

Sources checked in September 2026

This page is general parenting information and not medical or legal advice. Explore more in the AI for Kids hub.