Norway Says Kids First, AI Later
Norway just did something most countries won’t: it picked a side.
Generative AI is banned in elementary school classrooms starting this August. Ages six through thirteen — no ChatGPT, no Copilot, no Gemini scribbling their book reports. Ages fourteen to sixteen can use it, but only with a teacher watching. Seventeen and up? Go nuts, you’re old enough to decide.
Prime Minister Jonas Gahr Støre said it plainly: AI lets kids skip the hard parts of learning. Schools exist to teach reading, writing, and mathematics — not how to prompt a machine to do them for you.
This isn’t the reflexive Luddism people will call it. Norway already banned smartphones from classrooms in 2024, and the results are hard to argue with. Less bullying. Better grades. Fewer kids showing up at the psychologist’s office. The effects were especially strong for girls. That’s not theory — that’s data.
The “they need to learn AI” argument
This is the one I hear most: “AI is the future. If kids don’t learn it in school, they’ll be left behind.”
Sure. But you can’t learn to use a tool if you don’t understand what it’s replacing. Teaching a kid to write essays with ChatGPT before they can write an essay on their own is like teaching someone to use a calculator before they know what multiplication is. The calculator is faster. But the kid who never learned times tables can’t tell when the calculator is wrong. Same with AI. The kid who can’t write doesn’t know when the model is hallucinating. They can’t fact-check. They can’t spot the gap in the logic. They just trust the machine, because that’s all they’ve ever done.
The “this is just panic” argument
Every new technology gets a moral panic. Socrates worried writing would destroy memory. People freaked out about television, about video games, about smartphones. This is just the next one.
Maybe. But the smartphone panic turned out to be pretty well-founded. We have a decade of research showing that unrestricted device access in schools correlates with worse outcomes across the board. Norway’s own data confirms it. If the “panic” over smartphones was actually correct, maybe the “panic” over putting LLMs in front of eight-year-olds deserves a hearing too.
The counterargument I actually respect
Here’s the one that gives me pause: banning AI in schools means kids won’t learn how to spot AI-generated bullshit. If you take the tool away entirely, you lose the chance to teach critical evaluation alongside it.
I think there’s something to this. But the Norwegian model already handles it — you teach supervised use to 14-16 year olds. That’s exactly the right age to start a conversation about “here’s how this works, here’s what it gets wrong, here’s why you still need to know the material.” You don’t need to start that conversation at age seven.
What this really is
Norway is doing something boring and correct: treating AI as what it is. A tool with known failure modes, unknown long-term effects on developing cognition, and a commercial incentive to maximize engagement at the expense of learning.
The US is talking about the GUARD Act — age verification for AI companions, maybe. The language keeps getting softer. The carve-outs keep getting wider. Meanwhile, Norway just did the thing.
Kids need the fundamentals before they get the shortcut. That’s not anti-progress. That’s the opposite of anti-progress. A generation that can’t read, write, or do math without an LLM autocorrecting every thought isn’t the future we want. It’s the future we’re sleepwalking into.
Norway woke up.
Sources: Engadget via Reuters, covering the June 19 announcement and the 2024 smartphone ban results. Also discussed on Slashdot.