How to talk to your child about AI (when you're not sure yourself)

Here is the thing nobody tells you about talking to your child about AI: you do not need to know how it works. You just need to be willing to not know, out loud, in front of them.

Most parents who contact us say some version of the same thing: "I feel like I should be able to explain this, but I can't. And I'm worried about looking stupid in front of my own child." It is an entirely understandable feeling, and it is also a mistake — because the moment you pretend to know more than you do, you have already lost the most valuable thing about the conversation.

Your child does not need an expert parent. They need a curious one.

Start with what you have both already noticed

The best conversations about AI do not start with "let me explain how AI works." They start with something your child has already encountered — and probably already has a view about.

A few good starting points that work for almost any age:

• The autocomplete on a phone keyboard — why does it suggest the words it does?
• A streaming service recommending something new — how did it know they might like that?
• A chatbot they have used for homework, or heard about from a friend
• A voice assistant getting something noticeably wrong
• A photo filter that made someone look different

You do not need to explain how any of these work. You just need to ask: "Have you ever wondered how that decides what to show you?" That is enough to start.

The question is more useful than the answer

One of the most common parenting traps with technical subjects is the pressure to explain. A child asks how AI works, and the parent either wings a answer they are not confident about, or shuts down the conversation with "it's complicated." Neither is useful.

There is a third option: wonder out loud alongside them.

"I'm not completely sure how that works. What do you think is happening? Let's see if we can figure it out together."

This is not a failure of parenting. This is excellent modelling. You are showing your child that curiosity does not require prior knowledge, that uncertainty is something you sit with rather than paper over, and that adults do not always have the answers either. These are not small things.

Three conversations that are actually manageable

If you want something more structured, here are three conversations that work well for children aged 6–12, require no technical background, and naturally lead somewhere interesting.

Conversation 1: "Who decided that?"

Next time an algorithm does something unexpected — recommends a strange video, autocorrects something in a funny way, gets a search result obviously wrong — ask: "Who do you think decided this would happen?" Let them think about the fact that a person made choices, wrote instructions, and that those instructions can be wrong or unfair. You do not need to know what a training dataset is. You just need to help them notice that there was a human decision somewhere upstream.

Conversation 2: "Is that fair?"

AI systems often behave differently for different people. Face recognition that works better on some faces than others. Search results that vary by location. Spell-checkers that are better at some accents. When you notice something like this, ask: "Is that fair? Why do you think that happens?" You will probably not know the technical answer. But the question is the point — not the answer.

Conversation 3: "What would change if it was wrong?"

AI is increasingly used to make decisions about people — which job applications get through, which loan applications get approved, which students get flagged for extra support. Ask your child: "If a computer was making a decision about you, and it was wrong, what would you want to be able to do about it?" This is a genuinely hard question with no easy answer. That is why it is worth asking.

What to do when they know more than you

Some children — particularly in the 10–14 age range — will already know more about specific AI tools than their parents. They use them for school. They have watched dozens of hours of YouTube about them. They have opinions.

This is not a problem. It is an asset, if you position yourself correctly.

The useful move here is not to compete, but to reframe your role. You are not the person who explains how AI works. You are the person who asks: "But is that a good thing?" or "What happens to the people who get it wrong?" or "Who benefits from that?" Your lived experience — of how institutions fail, how promises don't always land, how technology has changed things before — is exactly what your child does not have yet. That is your contribution to the conversation.

A note on getting it wrong

You will say something about AI that is not quite right. You will simplify something too much, or be confidently wrong about something, or discover mid-conversation that you were thinking about it backwards. This is fine. What you do next is what matters.

The right move is to say: "Actually, I think I had that wrong. Let me think about it again." Not because children expect parents to be perfect, but because children need to see adults revise their thinking when new information arrives. That is one of the most important things you can teach them — and it costs nothing.

The longer game

Your child is going to spend their adult life in a world shaped by AI in ways neither of you can predict yet. The specific conversations you have this year are not the point. What you are building is a habit — the habit of noticing, wondering, questioning, and talking about it together.

A five-minute conversation on a Monday morning. A question at the dinner table. A moment in the car when something strange happens on someone's phone and someone says: "Actually, why does that do that?"

That is what AI fluency looks like at home. It does not look like a lesson. It looks like curiosity, practised regularly, together.