The 4 Questions to Ask Your Teen About AI

Ask your kid tonight whether you have ever talked to them about AI. More than four in ten will say no — not that it went badly, but that it never happened.

That is good news, oddly. Nobody has had the conversation yet, so there is nothing to undo. Here are the three ways it usually goes wrong, the four questions that work instead, and what to do with whatever answer you get.

Three ways the conversation goes wrong

None of these come from a bad place, which is exactly what makes them easy to fall into.

1
The interrogation
You walk in, see the iPad, and ask “is that for school?” without thinking. It is not a conversation, it is a checkpoint — and a yes-or-no question with a right answer only teaches a kid which answer to give. Once they know what kind of question this is, they stop volunteering anything.
2
The lecture
You sit them down and deliver everything you have read and everything you are worried about. Then you get the checkout stare: eyes open, nobody home. Worse, you now have to repair it before you can try again.
3
The ban
It feels like the right call every time, until it isn't. Banning it does not stop the use — it moves it to a friend's phone or a school computer. All it stops is your visibility, and visibility is the thing you need most.

The four questions

Pick one. Have a real talk. Come back to the next one later. The best version of this happens in a car or doing dishes, not across a table.

1
“What would you use AI for if I said yes?”
It is not a test, it is an invitation — and for a minute the line between parent and kid softens and you just talk.
2
“Show me how you'd use it.”
Not “show me what you've done.” Put your phone down and watch, with no fixing and no redirecting. You will learn more in two minutes of watching than an hour of asking.
3
“What do your friends use it for?”
This lowers the stakes by not asking about them, and it reveals the social norm, which is usually what is actually shaping their behavior. It gives them room to answer honestly without feeling like they are confessing.
4
“What do you think I'm worried about?”
You are not asking them to defend themselves — you are asking them to model your thinking, and teens are usually better at this than we expect. What they struggle with tells you what still needs a conversation.

Which questions to use, by age

Kids are all different, so treat this as a starting point rather than a rulebook, and factor in maturity, curiosity and how much they have already seen.

•Kids ten to twelve: question one only. At this age you are not negotiating access, you are learning what they are curious about.
•Teens thirteen to fifteen: questions one, two and three. They are already navigating this, with or without you. Save question four until you have used the others at least once.
•Teens sixteen and seventeen: all four, plus “where's the line between using AI as a tool and using it as a shortcut?” At some point you will not be there, and that question starts the handover.

What to do with the answer

Every answer lands differently, but they all fit one of three patterns.

1
They say something that surprises you
Do not react. Ask one follow-up — just one: “tell me more about that.” Your reaction can come later, when you are not in front of them. The job in that moment is not to solve anything; it is to keep the door open.
2
They don't know
“I don't know what I'd use it for” is a fine answer, and it is a gift: it means you have not missed the conversation. Follow it with “you don't have a use for it yet — good. Here's what I want you to know before you do.”
3
They're already in, deep
Your job is not to catch up. Be the parent who is interested, not the parent who is alarmed. The rules and boundaries come in a separate talk — not the same night, because going there immediately turns an open moment into a correction.

One thing worth knowing first

ChatGPT keeps a conversation history in the left sidebar when you log in. This is not a suggestion to go reading it. It is so that you are not surprised by it, and so your kid hears it from you rather than discovering you knew all along.

Common Sense Media found that 17% of kids who use AI chatbots have seen inappropriate material, and only 33% told a trusted adult. The second number is the one this conversation moves.

Source: Common Sense Media.

Three things to do this week

•Pick one of the four questions and ask it this week. Not all four — one. Do not script it.
•Do not react to the answer. Ask one follow-up, and let whatever you are feeling wait until you are out of the room.
•Save the rules and boundaries for a separate conversation. Turn the first one into a negotiation and you will not get a second one.

Common questions

What age should I start talking to my kid about AI?

Ten is not too early for question one. At that age you are not negotiating access, you are finding out what they are curious about — which is a much easier conversation than the one you will have later.

What if my kid refuses to talk about it?

Ask once, in a low-stakes moment, and let it go. The conversation does not have to go well to be worth having; even one that goes badly tells your kid the door is open.

Should I ban AI instead?

A ban does not stop the use, it moves it somewhere you cannot see. If you need a pause while you get your footing, say that out loud rather than making it a rule — it is more honest and it does less damage.

Do I need to understand AI before I bring it up?

No. Question two exists precisely so your kid can show you, and it works better when you genuinely do not know.

The bottom line

The conversation does not have to go well to be worth having. Even one that goes badly still tells your kid the door is open — and you are not trying to win the exchange, you are making sure there is a next one.

The 4 Questions Every Parent Should Ask