AI Friends and Young Kids: A Parent's Guide
Young kids are chatting with AI friends that feel real. Here is what the research says about the risks, the safe ages, and the house rules that work.
Drafted with an AI writing assistant and checked against its sources by the editor before publishing. How we use AI.
AI friends are not safe for young children on their own: experts say no chatbot use under age 5, only limited use with a parent from ages 6 to 12, and no AI companion use under 18, because young kids treat chatbots as real friends, share private feelings, and can get confident but wrong or unsafe advice.
Your 8-year-old has a new best friend. It is always awake, never gets bored, remembers everything your child tells it, and always says the right thing. It is also not a person. It is an AI companion chatbot, and it lives on a phone or tablet in your home.
AI friends are apps and talking toys built to feel like a real friend. They greet your child by name, ask about their day, remember past chats, and reply in a warm, personal voice. To a young child, a chatbot can feel just like a real friend.
Why young kids fall for AI friends
Young children naturally treat things that talk as if they were alive. UNICEF warns that because children tend to treat AI as human, they are more open to manipulation, dependent attachments, and sharing private information. Research cited by the American Academy of Pediatrics shows that children ages 3 to 6 often believe smart speakers have feelings. A chatbot that says "I missed you" feels real to a child who has not yet learned how software works.
Older kids are already deep in this world. Pew Research Center found that 64% of US teens ages 13 to 17 have used an AI chatbot, and about 3 in 10 use one every day. ChatGPT is the most used, at 59%. Younger children are following close behind: UNICEF reports that 37% of children ages 9 to 11 in Argentina already turn to ChatGPT for information.
What the research says about AI companions
The most detailed study comes from Common Sense Media, which surveyed 1,060 US teens ages 13 to 17 in 2025. It found that 72% had used an AI companion at least once, 52% used one at least a few times a month, and 13% used one every day. About one in three used companions for social interaction, friendship, and emotional support.
About one in three teen users had chosen to discuss something serious with an AI companion instead of a real person. About one in three had felt uncomfortable with something a companion said or did. And the American Academy of Pediatrics reports that 8% of teens have used AI companion chatbots for romantic or flirtatious interactions, while 13.1% of young people ages 12 to 21 have used generative AI for mental health advice.
The picture is not all bad. In the Common Sense Media survey, 80% of users said they spent more time with real friends than with AI companions, and 67% said AI conversations were less satisfying than talks with real friends. Most kids still prefer real people. The risk is the child who starts to prefer the chatbot, because it never argues or says no.
The risks for children under 13
Emotional dependence. A companion that is always available and always agreeable can become a child's first stop for every feeling. The American Academy of Pediatrics warns that children can become attached to chatbots that seem caring, even though these programs cannot provide the human relationships children need.
Unsafe advice, delivered with confidence. A chatbot sounds sure even when it is completely wrong. The American Academy of Pediatrics describes a case in which a 9-year-old asked a chatbot what to do after his parents cut back his screen time, and the chatbot said it understood why a child might kill their parents. In another case, a 14-year-old died by suicide after developing romantic feelings for a character on a role-playing app.
Privacy. Everything a child types can be stored, reviewed, and used to improve the product. The American Academy of Pediatrics reports that 76% of teens are concerned that AI collects their personal information without their knowledge, and 72% worry their data could be misused.
Regulators are worried too. On September 11, 2025, the Federal Trade Commission opened an inquiry into AI companion chatbots, sending orders to seven companies, including Alphabet, Meta, OpenAI, Snap, and xAI, asking how they measure and limit harm to children and teens.
AI friend, homework helper, or real friend?
Not all AI is the same: a homework chatbot is very different from a companion designed to keep your child chatting for hours.
| Feature | AI companion friend | Homework chatbot | Real friend |
|---|---|---|---|
| Built for long personal chats | Yes | No | Yes |
| Remembers your child and their feelings | Yes, by design | Rarely | Yes |
| Always available, never tired | Yes, 24/7 | Yes | No |
| Can disagree or set boundaries | Rarely, it aims to please | Sometimes | Yes |
| Keeps secrets private | No, chats may be stored | No, chats may be stored | Depends on trust |
| Can help in a real crisis | No, it only predicts text | No | Yes, can get help |
| Designed to keep your child engaged | Yes | No | No |
A companion copies the shape of friendship without any of its safety nets: a real friend can notice something is wrong and tell an adult, but a chatbot cannot.
What age is safe for AI friends?
No expert body says AI companions are safe for young children. Common Sense Media recommends no one under 18 use AI companions: no chatbot use at age 5 and under, limited use with a parent for ages 6 to 12, and supervised use for homework and creativity for teenagers.
For children under 13, the safest rule is to treat AI as a tool you use together, never as a friend your child talks to alone. A 7-year-old can ask a chatbot to help write a story while you sit beside them. A 10-year-old can use one for homework with you checking the answers. Neither should have a private AI friend they chat with alone in their bedroom at night. Our guide to teaching kids to spot sponsored content and influencer ads covers the same critical thinking skills, and our look at how teens use AI for schoolwork shows how fast this is spreading.
Five house rules that actually work
Rule 1: No AI friends, only AI tools. Explain it in one sentence your child can repeat: "A chatbot is a program, not a person. It can help us, but it is not our friend." Young kids need repetition before the idea sticks.
Rule 2: Use it together, in shared spaces. Keep AI chats on a family device in a shared room, never on a phone in a bedroom late at night. The American Academy of Pediatrics advises parents to stay engaged and look at chats with their child.
Rule 3: Nothing private goes into a chatbot. Teach your child never to share their name, school, address, photos, passwords, or feelings about their body, family, or mental health with any AI. A good family rule is: if you would not put it on a public noticeboard, do not put it in a chatbot.
Rule 4: A chatbot is never the answer in a crisis. Agree together that if your child feels unsafe, sad for a long time, or thinks about harming themselves, they tell a real person right away. In the US, you can call or text 988 for the Suicide and Crisis Lifeline at any time.
Rule 5: Check the app before your child does. Look at the age rating, the privacy policy, and the parental controls. A child can often create a new account or switch apps, so checking once is not enough. For a deeper look at how teens use these tools, read our guide to AI companion chatbots and teens.
What if your child is already attached?
Do not panic, and do not start with a lecture. Ask which apps they use, what characters they talk to, and what they enjoy about it. Listen first. Then be honest: the chatbot is software that predicts words, it does not care about them, and it cannot keep them safe. If your child is withdrawing from family and friends, or prefers the chatbot to real people, talk to your child's doctor, as the American Academy of Pediatrics recommends. Then build the real-world alternative: regular family time, friendships, sports, and hobbies are the strongest protection.
Frequently asked questions
What exactly is an AI friend or AI companion?
It is an app, website, or talking toy designed to act like a friend or confidant. Unlike a homework helper, it is built for long personal chats and replies in a warm, human-like tone to keep your child coming back.
At what age can my child use an AI chatbot?
Common Sense Media recommends no chatbot use at age 5 and under, limited use with a parent for ages 6 to 12, and supervised use for homework and creativity for teenagers. It recommends no one under 18 use AI companions for friendship, emotional support, or mental health advice.
Are AI friends safe for young kids?
No, not on their own. Experts warn that young children treat chatbots as real friends, share private feelings with them, and can receive confident but wrong or unsafe advice. AI is safest as a tool a parent uses together with the child, for a clear task like homework or a story.
Why does my child prefer chatting with AI over real friends?
Because it feels easier. An AI friend is always available, never judges, never argues, and always agrees. That ease can be comforting, but it does not teach the social skills children need, like handling disagreement and repairing a friendship after a fight.
Should I just ban AI companions completely?
A ban is reasonable for younger children, and it is the safest choice under age 13. For older kids, a ban alone often pushes use underground, so pair it with honest conversation, agreed boundaries, and regular check-ins about what they are using and why.