In the Age of AI, Should Kids Still Learn to Code?

At a free trial last month, a dad watched his son debug a robot for ten minutes, then turned to me and asked the question I now hear almost every week:

"With AI writing all the code now… is this even worth learning?"

It's a fair question. And I don't think the honest answer is the one you'd expect from someone who runs a coding school.

Here it is: the code was never the point.

If you want the full picture of what we do, start with our Programs and the free Parent Guide. But if you're weighing up whether coding still matters for your child in an AI world — this one's for you.

The short version (30 seconds)

Yes, AI can now generate code. Quickly, and often well.

But AI generates answers to problems someone else has already framed. It can't decide what's worth building, notice when the output is subtly wrong, or stay with a messy problem until it makes sense.

That judgement — framing the problem, spotting the error, knowing what "good" looks like — is exactly what a child practises when they build and code a robot that won't quite work.

So the skill that matters isn't typing code. It's the thinking underneath it. And in a world full of instant AI answers, that thinking is becoming more valuable, not less.

What this post covers

• Why "AI can do the coding" misunderstands the skill
• What AI still can't do — and why kids need to
• The abilities that get more valuable in an AI world
• Why ages 6–14 are the window
• Three conversations to have at home

"But won't AI just do the coding?"

For a lot of routine code — yes, increasingly.

I won't pretend otherwise. Tools that write code from a plain-English prompt are genuinely good, and they aren't going away.

But watch what actually happens when an adult uses one well. They decide what to build. They break the goal into parts. They read the result critically. They notice the bit that's wrong. They ask for a change. They test it.

Every one of those steps is a thinking skill — not a typing skill.

The person who can't do them doesn't get more powerful with AI. They just get faster at producing things they can't evaluate.

What AI can't do (yet) — and why kids still need to

Strip away the syntax, and coding is really four habits:

• Framing — turning a fuzzy goal ("make it follow the line") into precise steps.
• Sequencing — getting the order right, because the robot does exactly what you said, not what you meant.
• Debugging — comparing what you expected with what happened, then forming a theory.
• Judgement — deciding whether the result is actually any good.

AI is brilliant at filling in the middle. It's far weaker at the framing at the start and the judgement at the end — the human bookends. Those are the parts a child rehearses every single session.

Micro-story: "It works… but it's wrong"

An 11-year-old in one of our Caroline Springs sessions ran a block of code that worked perfectly — no errors at all. But her robot kept stopping just short of the line.

The screen said everything was fine. Her eyes said it wasn't.

"It works," she said slowly, "but it's wrong."

That sentence is the whole skill. A computer — and an AI — will happily run code that is technically correct and completely useless. Knowing the difference is a human judgement.

She made it at eleven.

The skills that get more valuable as AI gets better

When the "how" becomes cheap, the "what" and the "why" become the premium.

The abilities that compound in an AI world are the ones our sessions are quietly built around:

• Breaking a big, vague problem into small testable ones
• Reading a result critically instead of trusting it
• Staying calm and curious when something doesn't work
• Explaining your reasoning clearly — to a person or a machine

None of those have an expiry date. They're the same skills whether your child grows up to be an engineer, a nurse, a designer, or something that doesn't have a name yet.

It's why our slogan is what it is: the future is uncertain — life skills are not.

Why ages 6–14 are the window

Between 6 and 14, kids aren't just learning facts. They're forming beliefs about themselves:

• "I can figure hard things out."
• "Being stuck is the start, not the end."
• "I don't have to trust the first answer I'm handed."

That last one matters enormously in an age of confident, instant, occasionally-wrong AI. A child who's used to testing a robot's behaviour against what they expected is a child who will, later, test an AI's answer the same way.

We're not teaching kids to out-type a machine. We're teaching them to stay in charge of one.