Quick Answer
Start by asking, not informing. Ask a genuine question you do not already know the answer to: “Have you used AI for anything this week?” Then listen. For ages 6 to 9: build familiarity in small moments. For 10 to 12: do something together and discuss what you find. For 13 and up: skip safety and go straight to the judgment conversation.
How to talk to children about AI
Effective parent-child conversations about AI start with curiosity, not rules. Child psychologists and UNICEF both recommend beginning by asking what the child already thinks and uses AI for, rather than leading with safety warnings. The approach varies by age: small natural moments for young children, doing things together for 10 to 12-year-olds, and judgment-focused conversations for teenagers. Only 30% of parents are actively talking to their kids about AI use, despite 70% of teenagers using AI tools regularly.
Why Most Conversations Fail
Most parent-child AI conversations fail because they start with a conclusion. The parent already knows what they think, and the conversation is really just an announcement.
The first time I tried to have the AI conversation with my 13-year-old, I walked in with a plan and walked out having delivered a 10-minute speech to someone who was clearly waiting for it to end.
Your kid can tell the difference. It is the same tone that precedes conversations about phone use, social media, and not eating too much sugar. They switch into half-listening mode before you have said anything substantive. The goal of a good AI conversation is not compliance. It is understanding. You want to know what your kid actually thinks, what they are doing, and what they are confused or worried about. You cannot get that if you are doing all the talking.
How to Start
The best way to open a conversation about AI is to ask a genuine question you do not already know the answer to.
Not a test. Not a setup. An actual question. Some that have worked: “Have you used AI for anything this week? What for?” Or: “Did you see something about AI lately that was interesting or weird?” Or: “Someone told me AI can now do X. Does that seem right to you?”
The key is that you are genuinely curious. Kids have excellent radar for fake curiosity. My 9-year-old will talk about anything if she thinks I am actually interested. My 13-year-old will open up about things she would not normally mention if she senses I am not going to turn it into a lesson. The entry point matters more than you would think.
By Age
What the conversation looks like at each stage, with specific openers you can use this week.
Ages 6–9
Build Familiarity in Small Moments
You do not need a formal conversation. You need regular small moments that make AI a normal topic in your house, not a scary or forbidden one.
When AI comes up naturally — in a movie, at school, in a news story your kid overhears — pause and say something real. "That's AI. It's a computer that learned to do things by reading a lot of information." That is enough. You are building familiarity, not expertise.
At this age, the conversation is less about AI specifically and more about thinking for yourself. When my youngest brings home a question she already knows the answer to, I do not just confirm it. I ask her what she thinks first. That instinct to trust her own brain before outsourcing the answer is something I am actively building.
One thing worth saying to kids this age, directly and more than once: AI can be wrong. It sounds very sure of itself, but that does not mean it is right. That one sentence will pay dividends for years.
Conversation starters
"That's AI. It learned to do that by reading a lot of information."
"AI can be wrong, even when it sounds really sure."
"What do you think? Before we look it up."
Ages 10–12
Do Something Together
The most effective conversations with 10 to 12-year-olds involve doing something together and talking about what you observe.
Do not sit them down. Sit down with them. Open an AI tool. Ask it something you both care about. Look at the answer together. Then check it. Look up whether what the AI said is actually true. When you find something wrong — and you will — point at it and say: there it is. That is the thing to watch for.
This does two things. It makes AI concrete rather than abstract. And it models the exact behavior you want: skeptical curiosity. Not fear of AI, not uncritical trust of AI, but engaged, questioning use.
Some questions to work through together at this age: How could you check if this is actually true? Why do you think AI said it this way instead of another way? If you used this for your homework, what would you be adding yourself?
Conversation starters
"Let's ask it together and see if we think it's right."
"How would you check if that's actually true?"
"What would you add that AI didn't say?"
Ages 13+
Have the Judgment Conversation
By 13, shift from safety to strategy. The conversation that matters is about when AI makes you better and when it makes you dependent.
By 13, most kids are already using AI regularly whether their parents know it or not. The conversation that matters is not "should you use it." The conversation is about how and to what end.
The frame I keep coming back to with my oldest: tools are supposed to extend what you can do. A calculator extends what you can compute. But if you use a calculator so much that you lose the ability to do basic arithmetic in your head, the tool has made you smaller, not larger. AI is the same.
Ask: "Is there anything you used to be able to do yourself that you now feel like you need AI to do?" That question often gets silence first, and then honesty. Also have the emotional AI conversation directly. Many teenagers are using chatbots for companionship and emotional processing, not just homework. Ask what they get from talking to AI that feels different from talking to a person. You want to hear the actual answer.
Conversation starters
"Does AI make you better at this, or does it make the work disappear?"
"Is there anything you used to do yourself that you now feel like you need AI for?"
"What do you get from talking to AI that's different from talking to me or your friends?"
The Deeper Thing
Every conversation about AI is ultimately a conversation about your kid’s relationship with their own thinking.
Do they trust themselves? Do they know how to tolerate not knowing? Do they believe they are capable of figuring things out without help? Those are the questions AI surfaces. The tool did not create them. It just makes them visible.
The families that navigate this well are not the ones with the best AI policies. They are the ones where kids feel safe enough to say “I’ve been using AI for things I probably shouldn’t” without it turning into a big deal.
That kind of safety is built one small honest conversation at a time. For more on what to build before the AI conversation even starts, the Stoic Citadel covers psychological sovereignty, and the dependency guide covers what to watch for once AI use is already in motion.
This Week
One question. No agenda.
Pick the age-appropriate opener that fits your kid best. Ask it this week with no follow-up agenda — just listen to what comes back. You do not need to solve anything. You just need to open the channel and let them know you are paying attention without it feeling like surveillance.