How to teach kids critical thinking skills (with AI in the room)

Short answer: ask your child better questions than you answer, and make checking a habit rather than an exception. Critical thinking is not a subject you teach on a Saturday. It is the ordinary texture of how your family talks, and you can start changing that texture tonight with one sentence.

I have taught in classrooms for years and I have never once managed to install critical thinking in a child by explaining it. It arrives the same way an accent does: by living in a house where people talk that way.

Why this got urgent

For most of human history the clever child was the one who had memorised the most, and school was built around moving facts from a teacher's head into a pupil's. That model is finished. Your child now carries a device that produces a fluent answer to almost any factual question in under two seconds. Answers have become cheap. Questions have become priceless.

Chapter 5 uses an image I keep coming back to. An AI is a brilliant but literal-minded ship's navigator: it will calculate the fastest route to any destination and read you the currents, but it cannot decide where the ship should go. That is the captain's job, and the captain works by asking better questions. A child who consumes answers is being steered by the machine. A child who asks the questions is steering it.

Ask more than you answer

Here is the whole method in four words: answer questions with questions.

When my children ask me something factual my instinct, like yours, is to just tell them. It is faster and it ends the interruption. It also quietly teaches them that thinking is my job and receiving is theirs. Chapter 13 calls the alternative Practice One: make question quality a family value. The shift is from "what is the answer?" to "what is the better question?", and it costs about sixty seconds a day.

It feels awkward the first few times, for both of you. Push through it: what you are teaching is that their reasoning has value before any expert, or any machine, has weighed in. Two companion habits make it stick. Admit when you do not know, because "I have no idea, let's find out together" shows a child that not-knowing is a beginning rather than a failure. And at the school gate, ask "what was the most interesting question you asked today?" instead of "what did you get?"

The I Wonder Jar and Wonder Wednesday

The simplest structure for all of this sits on a kitchen counter. Label a jar or a notepad "I Wonder". Anyone in the family, adults included, who has a question they cannot answer writes it on a slip and drops it in. Profound or silly, both count. "Why do cats purr when they are in pain as well as when they are happy?" belongs in there. So does "why do we have eyebrows?"

Then, once a week, you pull a few out. In our house that is Wonder Wednesday. Ten minutes, phones in another room, and the rule is that everyone reasons out loud before anyone looks anything up. The point is not the right answer. It is practising uncertainty, and discovering that a good question is usually more interesting than any answer to it. Chapter 5 introduces the jar and Chapter 13 makes it Practice One. I have written about running both in full here.

Families who keep it going report the same thing: when their children eventually do reach for an AI, they ask it sharper questions, because they have practised forming the question before chasing the answer.

The ladder: Why, How, What if

Not all questions do the same work, and Chapter 5 sorts them into a rough ladder you can climb with your child.

Why? questions dig for causes. This is the phase every four-year-old runs on and every parent finds exhausting, and it is a real developmental milestone: the moment a child realises there is a world of cause and effect beyond what they can see. The best response is not an encyclopaedia entry. It is "that's a brilliant question. What do you think?"

The book's engineering version is the Five Whys game. Start with an observation, "the plant is turning brown", and ask why five times, each answer feeding the next question. Leaves dying. Not enough water. Dry soil. I keep forgetting. It lives in a room I never go into. Five questions from symptom to root cause.

How? questions dig for mechanisms, and they are the doorway to systems thinking. How does a car engine work? How did that apple get from a tree in another country into our trolley?

What if? questions open the hypothetical, and they are the foundation of every invention there has ever been. What if we could talk to animals? What if there was no gravity? What if we tried it the other way round?

Most families live entirely on the bottom rung. Modelling the top two out loud, in front of your child, expands what they think a question can be.

Critical thinking questions for kids, by age

The questions that work change as the child does. These are the ones I actually use.

Ages 5 to 8. Keep it concrete and sensory, and keep turning their question back to them.

• "What do you think?" (before anything else, every time)
• "How could we find out?"
• "Is that always true, or only sometimes?"
• "Shall we check whether that's right?"

Ages 9 to 12. This is the age for following a question to its source. Chapter 5 calls it the Junior Detective stage.

• "How do we know this?"
• "Who is telling us this, and why might they want us to believe it?"
• "What is this leaving out, and what would change your mind?"
• "If you had to argue the other side, what would you say?"

Ages 13 to 18. Teenagers can handle meta-curiosity: curiosity about their own thinking.

• "What did you believe six months ago that you now think was wrong?"
• "What would have to be true for that to be a good idea?"
• "Who benefits if everyone believes this?"
• "How confident are you, honestly, and what is that based on?"

That last band is where Chapter 5 introduces epistemic humility: clever people are wrong all the time, certainty is often a warning sign, and the best thinkers are the ones most willing to revise. It is the direct antidote to swallowing an AI answer whole.

V.E.T.: critical thinking applied to a machine

All of the above is general. AI needs one specific tool, because AI writing is fluent, structured and delivered in exactly the same confident tone whether it is right or wrong. That tool is the V.E.T. rule from Chapter 10, and it is covered in full here.

Verify. Can I find a second reliable source saying the same thing? Teach older children the tell: when an AI produces a very specific claim, a statistic, a date, a quotation, a named study, that is exactly when to check, because specificity is what these machines fake best.

Explain. Can I say this in my own words with the screen closed? If not, I have not learned it and it is not mine.

Transform. What have I added? My own example, my own view, the part I actually find interesting.

Underneath all of it sits the sentence I say most often at home: "Is this helping you think, or thinking for you?" Say it lightly, say it often, and eventually you will not have to say it at all, because it will be running in their head instead of yours.

Questions parents ask

At what age can you start teaching critical thinking?

From about five, as a game. At that age it is "what do you think?" and "shall we check whether that's true?", nothing more complicated. From nine you can start tracing claims back to sources, and from the early teens a child can examine their own reasoning and admit when it was wrong. The habit is the same at every age; only the difficulty changes.

What are good critical thinking questions for kids?

The best ones cannot be answered with a fact. "How do we know this?", "What is this leaving out?", "Who benefits if everyone believes it?", "What would change your mind?" and "If you had to argue the other side, what would you say?" all force reasoning rather than recall. Keep three or four in your back pocket for the dinner table.

Does using AI make children worse at thinking?

It depends entirely on how it is used. A child who pastes an AI answer has skipped the thinking; a child who argues with it, checks it and rewrites it has done more thinking than a worksheet would have demanded. The tool does not decide the outcome. The habit around it does, which is why the V.E.T. rule matters more than any ban.

How do I get my child to stop asking me for the answer?

Stop giving it, warmly. "Interesting. What do you think first?" then sixty seconds of silence. Expect resistance for a week or so, because you have changed the deal. Pair it with admitting your own ignorance out loud, so it does not look like a test you are setting them. Once they realise the reasoning itself is what you are interested in, most children start enjoying it.

Sources

• Chapter 5 of Raising AI-Ready Children on curiosity, the "I Wonder" Jar, the Five Whys game, the Why/How/What if ladder, and questioning AI outputs.
• Chapter 13, Practice One, on making question quality a family value and the "What do you think?" upgrade.
• Chapter 10 on the V.E.T. rule, the 20-Minute Rule and the Family AI Agreement.
• Gruber, M. J., Gelman, B. D., and Ranganath, C. (2014). States of curiosity modulate hippocampus-dependent learning via the dopaminergic circuit. Neuron, 84(2), 486 to 496.