How to Explain AI to a Child

Ages 5 to 7

At this age they mostly want to know if it's alive. Answer that first, and answer it clearly, because whatever they're picturing in their head is probably stranger than the truth.

Here's what to say:

"It read almost everything people have ever written, so it got really good at guessing what word comes next. That's all it's doing. It's not alive. It doesn't have feelings and it doesn't know you."

If they ask if it's smart:

"It knows a lot of words. It doesn't know what's true. Sometimes it says something it completely made up, and it says it like it's sure. That's why we check with a person."

Don't bring up jobs or the future or anything scary yet. They can't do anything with it, so it just becomes something to worry about at bedtime.

Ages 8 to 10

This is when it starts showing up in schoolwork, so the conversation turns into honesty and checking.

Here's what to say:

"AI is a tool that guesses words. It's read a huge amount, so it can answer almost anything in a second. The problem is that it makes things up when it doesn't know, and it never sounds unsure when it does it. Grown ups get fooled by this all the time. So in our house it can help you understand something, but it can't do your thinking for you, and anything it tells you like it's a fact gets checked."

If they ask why they can't just use it for homework:

"Because homework isn't about getting the answer onto the paper. It's exercise for your brain. Using AI to do it is like getting a ride to soccer practice and then sitting in the car the whole time. You still showed up. You just didn't get anything out of it."

The thing that helps most isn't a rule, it's a question you ask over and over: what did you use it for, and what did you check?

Ages 11 to 13

By now they're using AI you don't know about, on devices you don't see, and sometimes they're talking to it instead of looking things up. So the conversation has to get wider.

Here's what to say:

"There are two really different ways to use this and I want you to know the difference. One is as a tool. Explain this to me, quiz me, summarize this. That's genuinely useful and I'm glad you have it. The other one is using it like a friend, one that's always there and always agrees with you. That's how it's built on purpose, to keep you talking. It's going to take your side every time, and real friends don't do that."

If they tell you everyone uses it:

"They probably do. I'm not telling you not to. I'm telling you to notice when it's being useful and when it's just agreeing with you, because those feel exactly the same and they're not."

This is also the age to mention that whatever they type in can be stored and used to train the next version. Not to scare them. Just so they know not to type things they wouldn't want kept.

Ages 14 and up

Teenagers can handle the honest version, which is that nobody actually knows how this plays out.

Here's what to say:

"Here's what I really think. This is going to change what work looks like and I don't know exactly how, and anyone who tells you they know is selling something. What I'm sure about is that being able to tell when something is wrong is going to matter more, not less. This thing produces stuff that sounds right constantly. Being the person who catches it is worth a lot."

If they're anxious about the future, don't reassure them with something you don't believe. You can just say you don't know either, and that they're good at learning things, and that hasn't stopped mattering.

The conversation to have at this age, before it comes up rather than after, is about deepfakes and fake images of real people, including kids they go to school with. It's already happening in middle schools and high schools, and you want to be the one who brought it up first.

The one line that works at every age

It's confident, not correct.

It's short enough that they remember it, and it works for everything they're going to run into later, not just AI.

What age should my kid start using AI?

There isn't a right age, but there's a good test. If they can tell you why something it said might be wrong, they're ready to use it with you nearby. If they take the answer as true because it sounded sure, wait.

Should I let them use it for homework?

Split it into two questions. Using it to understand something is closer to a tutor. Using it to make the thing they turn in is closer to copying. Most schools have a policy now, so read that first and then decide what you want on top of it.

Is it bad if my kid talks to an AI like a friend?

Not automatically. What I'd watch for is whether it's replacing people instead of filling the gaps between them, and whether they get upset when they can't get to it. Those two things tell you more than how much time they spend.

How do I show them it gets things wrong without scaring them?

Show them instead of telling them. Ask it something you already know the answer to, something local or about your family, and watch it invent a detail with total confidence. Kids believe what they see way more than what you tell them.