The V.E.T. rule: teach your child to check what AI tells them

Short answer: never let your child treat an AI answer as a fact. Treat it as a suggestion to be checked. The way I teach that at home is three letters on the fridge: V.E.T. Verify it against another source. Explain it in your own words. Transform it into your own work. A child who runs every AI output through those three steps cannot be fooled by a confident machine, and cannot hand in something they do not understand.

Here is why it matters, how to teach it at each age, and what it looks like when it works.

The problem the rule solves

AI writing looks finished. It is fluent, structured, and delivered in exactly the same confident tone whether it is right or wrong. That is the trap. These tools do not know things; they predict words, and some of the time they invent facts, dates, quotations and even sources that do not exist. The industry calls it hallucination. I call it the reason my children are not allowed to paste.

Adults are fooled by this regularly. A child who has grown up with a machine that always sounds sure has no reason to doubt it unless someone teaches them to. That someone is you, and the teaching is simpler than it sounds.

V is for Verify

Is this actually true? Can I find two or three other reliable sources that say the same thing: a textbook, a reputable news site, a library book, the teacher?

For a young child this is a game. "The AI says giraffes sleep for twenty hours a day. Shall we check?" They will enjoy catching it out, and they will remember that it can be caught. For a teenager it is a habit that will serve them in every job: never cite a source you have not opened.

The tell to teach older children: when an AI gives a very specific claim, a statistic, a quotation, a study, that is exactly when to check, because specificity is what the machine fakes best.

E is for Explain

Can I explain this in my own words, laptop closed? If not, I do not understand it, and it is not mine.

This is the step that does the most work, because it converts reading into learning. Rephrasing something forces the brain to process it. A child who can explain the causes of the First World War conversationally over dinner has learned them. A child who can only read them off the screen has not, however good the paragraph looks.

It is also the quiet enforcer of honesty. You do not need detection software. You need to ask "so what did it say, then?" and listen.

T is for Transform

How do I make this my own? What can I add: my own example, my own view, my own voice, the thing I actually find interesting about it?

The AI's output is the starting point, never the finishing point. Transforming it is where your child's thinking shows up, and it is the difference between work a teacher recognises as theirs and work that sounds like everyone else's homework that week.

By age

Ages 5 to 8. Only V, and as a game. "Let's check whether that's true." The concept you are planting is that the machine makes things up and is not always right. That single idea is the foundation of everything later.

Ages 9 to 12. V and E. Verify against one other source; explain it back to you. Put the three letters on the fridge anyway, so the third is waiting when they are ready.

Ages 13 to 18. All three, and increasingly on their own. The test I hand over at this age is "could I explain this to someone without looking?" If the answer is no, they know what to do before I say anything.

Why the rule comes with a reason

Every rule in my house has a why, because rules with reasons survive when I am not in the room. V.E.T. is not "because you'll get caught". It is "because we value truth, and we think for ourselves instead of taking things at face value". Say that in your own words until your child can say it back. A teenager who believes it will check what an AI tells them at university, at work, and in the voting booth.

Questions parents ask

What is the V.E.T. rule?

A three-step check for anything an AI produces: Verify it against a second reliable source, Explain it in your own words, and Transform it into your own work with your own ideas. It comes from Chapter 10 of Raising AI-Ready Children and is designed to be stuck on a fridge.

Does AI really make things up?

Yes. Language models predict words rather than look up facts, and they sometimes invent statistics, quotations and sources, delivered in the same confident tone as the truth. It is common enough that checking should be a habit, not an exception, especially for specific claims.

At what age can a child start checking AI answers?

From about five as a game: "shall we check whether that's true?" The full three-step rule suits children from around nine, and by the early teens it should be something they do without being asked.

How do I check if my child understood, not just copied?

Ask them to explain it with the screen closed. A child who can talk you through the work understands it; a child who cannot, does not, whether or not an AI was involved. It takes two minutes and needs no software.

Sources

• Chapter 9 of Raising AI-Ready Children on how language models work and why they hallucinate.
• Chapter 10 on the V.E.T. checklist, the 20-Minute Rule and the Family AI Agreement.