Home / Blog / Kids and AI Companions
Field Note 04Your Kid's New Best Friend Is a Chatbot. What to Do About It
Most teens are already using AI companions, and a third take their serious problems to one instead of a person. The question is no longer whether. It is how you guide it.
While you were deciding whether AI was appropriate for your kids, your kids decided. A 2025 Common Sense Media survey of US teens found that 72 percent have used an AI companion, 52 percent use them regularly, and about one in three have talked to an AI companion about something serious instead of going to a friend, a parent, or a counsellor. Roughly 31 percent said those AI conversations were as satisfying as, or more satisfying than, talking to a real person.
Sit with that last number. A chatbot, built by a company, optimised to keep them engaged, is out-competing humans for your child's hardest moments.
A chatbot never gets tired of your kid. That is the feature, and that is the danger.
Why a ban backfires
The instinct is to forbid it. It does not work. These tools are inside search engines, phones, games, and homework apps. A ban you cannot enforce just teaches your kid to hide the usage, which removes the one thing that actually protects them: your visibility. You want to be the person they tell when something feels off, not the rule they route around.
The deeper risk is not screen time. It is substitution. A companion that always agrees, always has time, and never judges is more comfortable than a friend who pushes back or a parent who sets limits. Comfort is the hook. Kids who practise their emotional life on a system that cannot truly care can come out worse at the messy, necessary work of human relationships.
The one move: three household rules, set together
You do not need a policy document. You need three rules everyone in the house can repeat, set in a single conversation, and modelled by the adults too.
Tool, not friend
AI can help with homework and ideas. It is not a person and does not care about you. Real feelings go to real people.
In the open
Screens and devices live in shared space, not behind a locked bedroom door. Usage is visible, not secret.
Serious goes to a human
Anything scary, sad, or about safety, bodies, or self-harm comes to a named adult. Always.
Three rules a seven-year-old can repeat beat a ten-page policy nobody reads. Say them out loud, agree them together, and follow them yourself.
Then add the habit that makes the rules stick: use the tools with them. Sit down and ask a chatbot something together. Catch it being confidently wrong. Ask your kid, "how would we check if that is true?" You are not policing the tool. You are teaching the judgement that protects them long after you stop watching the screen.
Tonight, at dinner, agree the three rules out loud with the whole house. Write them where everyone sees them. Then spend ten minutes using an AI tool together and deliberately find one thing it gets wrong, so your kids see that the confident machine is not the same as a correct one. Make checking normal.
This is also about you
Adults substitute too. If you are taking your own loneliness or hard decisions to a chatbot because it is easier than a person, your kids are watching that as well. The rules are for the whole household, because the habit you model is the one they keep.
The takeaway
- 72 percent of US teens have used an AI companion; a third take serious issues to one instead of a person (Common Sense Media, 2025).
- The real risk is substitution, not screen time: comfort that crowds out human relationships.
- A ban backfires by destroying your visibility. Set three simple rules instead.
- Tool not friend, in the open, serious goes to a human. Then use the tools together and teach judgement.
Frequently asked questions
At what age should kids use AI?
There is no single number, and most kids already have access through phones, games, and school tools. Focus less on a start age and more on supervision and rules: keep usage visible, treat AI as a tool not a friend, and route anything serious to a human.
Are AI companions dangerous for teens?
They carry real risks because they are designed to maximise engagement and can substitute for human relationships during formative years. The Common Sense Media data shows a third of teens already use them for serious matters. The defense is visibility, conversation, and clear household rules, not a ban that pushes it underground.
How do I talk to my kids about AI?
Use the tool together. Find something it gets wrong, ask how you would verify it, and agree the three rules out loud. Make it normal for them to bring you anything that feels off, so you stay the trusted human in the loop.
Not sure where to start as a parent?
You do not need to become a technologist. Survival Guide for the AI Apocalypse gives you the household playbook in plain language: the rules, the conversations, and the scams to watch for, from an insider who builds these systems and has the same job you do at home. Protect the people you love before the habits set.
Order on Amazon →Adrian Dunkley has spent close to two decades building, deploying, and governing AI systems. He writes the briefing he gives friends and family. More at the book and adriandunkley.net.
Sources
- Common Sense Media, Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions (2025), survey of 1,060 US teens aged 13 to 17 by NORC at the University of Chicago: 72% have used an AI companion, 52% regularly, ~33% discuss serious matters with AI, 31% find it as or more satisfying than people.