Norway is, by almost any metric, a profoundly successful nation. It’s rich, democratic and relatively corruption-free. It’s not a socialist country, but fans of a robust welfare state and high taxes see much to admire in the very progressive Norwegian model. It also benefits from having the biggest and arguably best-run sovereign wealth fund in the world.
And yet, Norway nearly ruined its children.
In 2016, flush with cash and progressive values, Norway gave every child in the country, starting at the age of 5, their own iPad or similar digital device. A decade later, young Norwegians now struggle to read. “Around 500,000 Norwegians, in a population of only 5.6 million, cannot read a text message or simple instructions,” reports The Times of London. “Of the 65 countries measured for children’s enjoyment of reading by Pirls (Progress in International Reading Literacy Study), it comes bottom.”
Norwegian Prime Minister Jonas Gahr Store launched a program in August to deal with the problem. “Norwegian children used to be among the best readers in the world,” he said at the time. “But today, 15,000 pupils finish primary school without being able to read properly.”
Now imagine giving every child an AI chatbot to answer all of their questions.
I am not a catastrophist when it comes to artificial intelligence. But, given Norway’s experience with iPads — or our own with smartphones — I worry that the mass introduction of AI, particularly in schools, will be very bad for children.
As a curmudgeon and as a writer, I hate nearly all of the cliches about children, even the ones that are accurate. With that caveat, it’s simply true that children are the future. They will be the next generation of parents, voters and citizens. And all of the cliches about how kids learn by doing are true.
AI removes the doing.
Just as you can’t learn how to ride a bike from reading a book, you can’t get the benefits of reading by asking AI to read a book for you. The same holds for math, science, computer programming and nearly every other aspect of education.
Our military is the finest and most lethal in the world. But before you learn how to operate a drone or launch a cyberattack, you still have to go through basic training.
Education, both at home and at school, is basic training for civilization.
Americans love technology, but not every technological advancement is an advancement in every sphere of life. There are machines that can lift weights, but using a machine to lift weights for you doesn’t count as exercise and doesn’t build your muscles. Using a machine to think for you is the route to mental flabbiness.
Fans of AI don’t like this argument. They use terms like “cognitive shift” and “upskilling.” By removing the drudge work, smart people can use AI to be smarter and more productive. I think there’s a lot of merit to this when talking about existing highly skilled workers. But how did those workers become highly skilled in the first place? By doing the work.
Educators learned a similar lesson when cheap calculators became widely available in the 1970s. As you got more advanced, you could use calculators for certain problems. But first you needed to learn how to do the basics. You also needed to learn to think mathematically. AI is essentially a souped-up calculator for nearly all mental tasks. Again, that’s often — but not always — great for adults who grew up in a pre-AI world.
Which is why I think education should mostly stay in the pre-AI world. That will be very difficult. It will require more memorization, more tests in the classroom and an education establishment that can resist the seduction of technological fads. If the point of education is to build up muscle memory for how to think and how to do things, letting kids go home and have AI give them the answers is not very different, educationally, from letting kids cheat. For the same reason that having a robot do 50 push-ups for you wouldn’t be acceptable for a physical fitness test, having a robot read a book for you shouldn’t be acceptable either.
The point of education in the AI era shouldn’t be to teach kids how to find the answers in the most efficient way possible, but to equip them to be ready to ask the right kinds of questions, including the ability to ask an AI chatbot why it gave you the answers it did.
Jonah Goldberg is editor-in-chief of The Dispatch and the host of The Remnant podcast. His Twitter handle is @JonahDispatch.
