When I was a kid, my dad and I often talked about a story we wanted to write together. It was called “The Day Nobody Went to Disneyland.” The idea was that one day, the weather is so perfect that everyone assumes Disneyland will be packed, so they all decide to stay home. But my dad, thinking the same thing, sees how disappointed I look that morning—and decides to take the chance anyway. We head off, and when we get there, the park is completely empty. No one else was bold enough to show up. We get to enjoy everything—the rides, the shows, the food—all to ourselves.
I loved the idea of writing that story. But, as often happens, life got in the way. We were too busy to make it happen, time slipped by, I grew up, and we both forgot about it.
Thirty years later, I was at my parents’ house for dinner with my own kids (ages eight, five, and two). For some reason, the idea popped back into my head. I told my oldest about it, and she wanted to hear the whole thing. So we turned to ChatGPT, gave it a rough outline, and asked it to write the story in the style of Dr. Seuss (one of our favorites). Within seconds, we had it. She read it out loud, and we laughed.
Of course, what the AI produced wasn’t perfect. Some lines didn’t quite make sense, and a few Americanisms had slipped in—like “sneakers” and “cinnamon buns.” But that night on the couch, we had a great time tweaking the prompts, refining the story, adding more scenes, and working with this strange new creative partner to do something that had slipped through my fingers for decades.
For the past 15 years, I’ve been studying how AI affects work and society. But becoming a dad to three young kids has made these changes feel much more personal: it’s clear their future will look wildly different from my past.
For my wife, that realization really hit home a few weeks ago during a drive to her parents’ place in Suffolk. Three kids stuck in a car, crawling along the A12, is a recipe for chaos. So we were thrilled to find a rare podcast that everyone actually enjoyed: History’s Not Boring, a series of 15-minute conversations on all sorts of topics, narrated by two kids.
As we listened, we started guessing who these clever kids might be. Where were they from? How were they chosen? How did they balance this with school? Eventually, I looked them up online. Every guess we had was wrong. The kids didn’t actually exist. The whole podcast seemed to be made by AI. None of us had realized.
It was a strange moment, finding out that people we’d grown fond of weren’t real at all. It was also impressive that AI could create such high-quality educational content. But for my wife, who makes podcasts and documentaries, it was unsettling too: here was something that would have taken her and a talented team days to produce, now being done without any people involved.
My work on AI has taken me to boardrooms, conference halls, classrooms, and government buildings. But wherever I go, one question comes up more than any other: what should my children do?
Right now, we’re all a bit lost. Teachers worry that their old methods don’t work anymore. Parents, watching their kids use AI to solve problems and answer questions, wonder what they actually know. Employers doubt whether traditional academic achievements—earned through coursework, exams, and assessments—still tell us anything useful about what young people can do.
But reacting to these new technologies by putting up walls, as many instinctively want to—banning, scolding, punishing—just can’t be the right answer. For starters, this is the world the next generation will live in: we’re failing them if we prepare them for the world we grew up in, rather than the one they’ll actually inhabit.not the one they will actually inhabit. How often, for example, have you heard people accuse students of cheating if they use AI to help with their work? But at what point are we cheating them by failing to rethink how we educate them? If AI is making education “too easy” for young people, when does the responsibility shift to us, the adults, to make their learning more challenging, to push them further, to stretch their abilities?
More importantly, this dismissive attitude toward AI is also unimaginative. It stops us from seeing the amazing possibilities AI could create for the next generation. Of course, there are risks. But at the same time, if we use this technology wisely, what difficult ideas might the next generation now grasp—ones that were beyond us when we were their age? What hard problems might they solve—ones that once required years of training and experience, if they could be solved at all?
Forget future-proofing skills
A big part of the challenge with AI is that our traditional response to technological disruption in the working world—the idea of “future-proofing” people—no longer works. In 2013, the then prime minister David Cameron announced that England would become the first place in the world where all primary and secondary school children would learn to code. This, according to education secretary Michael Gove, would “equip every child with the computing skills they need to succeed in the 21st century.”
The idea seemed bold and clever: in the years that followed, it was hard to find an advanced country that didn’t follow our example.
Fast-forward to today. What do the latest AI systems, such as ChatGPT and Claude, turn out to be best at? Writing code. In January 2026, Anthropic reported that 90% of the code for Claude Code—their AI-powered coding assistant—was itself written by AI. Skills that were meant to protect kids from technological disruption for their entire lives were largely redundant before they even left school.
We must use AI critically, not blindly, keeping our basics sharp so we can tell when AI is being a savant—or an idiot.
For politicians and policymakers, it’s tempting to dismiss the coding saga as an unfortunate blip. But this is a mistake. The reason they slipped up was because they had believed in the idea of “future-proof” skills—the thought that there exists some valuable set of skills that AI won’t be able to do for some time, and that by thinking deeply enough about the future, we can accurately identify them.
The truth is that we know only two things about what lies ahead. One is that it will be full of technologies far more capable than today’s. The other is that we know little else. But rather than trying to resolve this uncertainty, we have to accept it and instead ask a different question: how do we prepare the next generation to thrive in a future we actually know surprisingly little about?
Get back to basics
The first step is to get back to basics. Since 2009, literacy and numeracy have been falling among young people around the world, according to the OECD’s programme for international student assessment (Pisa); the same holds true for adults. On its own, this is worrying. But given the uncertainty we face about the future, these trends are a disaster.
Why such a calamity? To begin with, while it might be hard to say precisely which more advanced skills will turn out to be most valuable in the future—creativity, judgment, something else—they will rely in some way on these basic skills. They are the foundations for everything else.
What’s more, AI systems are far from flawless. They make mistakes, often on simple problems; they hallucinate, providing confident and plausible answers that are completely made up. They are, as the computer scientist Geoffrey Hinton put it, “idiot savants.” That means we must use AI critically, not blindly, keeping our basics sharp so we can tell when AI is being a savant—or an idiot.We can tell when AI is acting like a genius—or a fool.
With both of those reasons in mind, we should be putting a lot of effort into teaching reading, writing, and math, even if AI seems to do them better, which it increasingly does. Going back to the basics is what an economist would call a “no-regrets” strategy—we will never regret improving these core skills, no matter how the future plays out.
I have used AI to answer my five-year-old’s bedtime-delaying habit of asking huge questions just as I turn off the lights.
And we should try to be creative. When my eight-year-old got bored learning her times tables the old-fashioned way—long lists, memorized by repetition—I turned to ChatGPT to help her design various computer games that tested what she knew. (Unicorns and rainbows showed up in every version.)
Then there was the time my five-year-old was tired of reading his standard phonics books after school—there’s only so much Biff, Chip, and Kipper a person can take—so instead, we used AI to create some custom stories together, carefully matched to his reading level and much closer to his favorite topics. (At the time of writing, that meant HMS Belfast, TNT explosives, and Bukayo Saka.)
Trying new things matters, and we can learn a lot from what others come up with. Take Chris Moran, for example, head of editorial innovation at the Guardian. His daughter had been reading Dracula at school but was struggling to connect the sprawling novel to the real world. Together, with the help of AI, they built an app called PlotLines, which placed the story on an interactive 1890s Ordnance Survey Map, mapping out the key scenes and character journeys across Europe with AI’s help. (They’ve since done this for many other books too.)
These are exactly the kinds of innovations we should be testing and welcoming throughout education—not banning.
View image in fullscreen
Teach both, test both
Even with all the uncertainty we face, there’s still one thing we know about the future: it will be full of technologies far more powerful than today’s. With that in mind, we need to teach the next generation how to use them. Doing this properly will require us to set aside a serious amount of time for teaching people how to use AI.
Skip past newsletter promotion
Free newsletter | Weekly
Sign up to Inside Saturday
The only way to get a behind-the-scenes look at the Saturday magazine. Sign up to get the inside story from our top writers, plus all the must-read articles and columns, delivered to your inbox every weekend.
Enter your email Sign up
After newsletter promotion
The challenge, though, is how to do this without also making them forget important basic skills. A growing concern is that AI is making us less smart. Why read a book if AI can summarize it? Why finish math homework if AI will do the calculations? In my view, the answer doesn’t come from Silicon Valley, but from a forgotten British math professor, Wilfred Halliday Cockcroft, half a century ago.
In the 1970s, the quality of math teaching in UK schools seemed to be falling apart, and numeracy levels were reportedly low. Cockcroft, who had a long-standing interest in math education, was asked by the government to look into it. In 1982, the Cockcroft Report was published. It was massive, detailed—and now mostly forgotten. But it might turn out to be the most important document for thinking about the future of education, because of how it dealt with the electronic calculator.
‘I thought, I’ve tried everything else, why not give AI a shot?’: the long-lost family reunited by ChatGPT
Read more
It’s fascinating to read Cockcroft’s report and see how the challenges of the calculator back then were so similar to the ones we face with AI today—from the fear that it would weaken “basic skills” to the intimidation many felt facing this strange new technology (“Some … had been discouraged by theA large number of figures had appeared after the decimal point. Cockcroft was realistic: he expected that “all candidates will have access to a calculator by 1985.” His proposal was revolutionary: he wanted to overhaul mathematics education by splitting it into two parts—one part teaching students how to use a calculator, and the other part teaching them how to manage without one. Crucially, both skills would be tested. The idea caught on. Today, learning math both with and without a calculator is the global standard. The first approach builds basic skills, while the second applies them to more challenging problems.
We should take the same realistic and revolutionary approach with AI now, using this proven principle—what I call “teach both, test both.” Every subject, from history to English literature, should be divided into two parts: one part teaching students how to use AI, and the other teaching them how to succeed without it. Both should be examined. A teacher cannot monitor whether a student uses AI in the privacy of their own room. But nothing can replace the experience of sitting in an exam, looking at the paper, and feeling that cold sweat when you realize you haven’t prepared for both sections.
Not all screens are bad. One of my concerns is that we let a legitimate worry about social media’s impact on children’s lives spill over into how we think about AI. This could turn sensible restrictions on social media into knee-jerk bans on AI. But social media and AI are not the same thing. While social media often dehumanizes us and pulls us away from the real world, AI—when used properly—can greatly improve our lives.
There is another way to think about this. A debate is currently underway about the “Goldilocks” amount of screen time for children—not too much, not too little, but just right. However, this is the wrong debate to have. What matters is not so much how much time is spent on screens, but what is actually on those screens and what we are using the technology to do.
This distinction matters not only for what we teach, but also for how we teach. Consider another example: personal tutoring. It is often said that an average student who receives one-to-one tutoring will outperform almost all their peers in a traditional classroom. I saw this firsthand during my years as a tutor in Oxford, teaching mathematics and economics.
The problem is that human tutors are too expensive to provide for everyone. But now, AI can finally offer high-quality personal tutoring. It can tailor the way material is taught to each student’s unique strengths and weaknesses, mimicking interactions with a human tutor but at a much lower cost.
Honestly, AI provides a level of personalized instruction that I—and many other teachers I have observed over the years—have struggled to achieve. Part of what is striking is AI’s breadth. I have used it to respond to my five-year-old’s bedtime-delaying tactic of asking huge questions just as I turn off the lights. When he asked, “Daddy, where did the first human come from?” AI created a short story about evolution. I have also used it to generate step-by-step instructions to help graduate students solve difficult mathematical problems in economics, like the Ramsey growth model.
“I don’t think anyone has ever paid such pure attention to me and my thinking and my questions,” a student was reported as saying in The New Yorker in April 2025. “It’s made me rethink my interactions with people.” AI never gets tired or distracted. It doesn’t have office hours or limited classroom time. It will always answer one more question and always provide another explanation of a problem you don’t understand.
Focus on problems, not the job. I am not suggesting…It’s no surprise that students have been booing tech leaders at recent US graduation ceremonies. Over the past 15 years, the most resistant groups I’ve spoken to are young professionals. They’ve invested a huge amount of time and money into becoming lawyers, doctors, or whatever they’ve chosen, and they’re understandably angry when, just as they’re about to finish, they’re told the world they’ve been preparing for no longer exists.
There’s so much advice I’d want to give them as they start their careers, but one of the most important pieces is this: choose a profession because the problem interests you, not the job itself. To put it bluntly, if you go into medicine because you like the doctors on House, or law because of Suits, or marketing because of Mad Men, you’ll be disappointed. These jobs are about to change dramatically.
But the problems themselves—improving health outcomes, offering legal advice, selling products—aren’t going away. What will change is how we solve them and the skills we need to do so.
You could see this coming long before generative AI. Back in 2017, a team at Stanford University announced they’d built a system that could tell from a photo whether a freckle was cancerous, just as accurately as top dermatologists. It was a breakthrough moment in AI-powered diagnosis. But what stood out most was that the final author on the Nature paper announcing this was Sebastian Thrun—not a doctor, but a leading computer scientist who created the world’s first driverless car. Here was someone with almost no medical knowledge, yet his skills allowed him to build a system that matched the best doctors.
Run toward AI and science
If I were starting my career today, at the very beginning, I’d run toward AI and science without hesitation—not because they’re safe from automation, but because that’s where the most exciting developments will happen in the coming years. In the 20th century, our best ideas came from brilliant human minds. In the 21st, I expect they’ll come from capable AIs instead.
We got a glimpse of this in late 2024, when the creators of DeepMind’s AlphaFold2 won the Nobel Prize in Chemistry for cracking the “protein folding problem”—one of biology’s biggest unsolved challenges, key to understanding diseases and how to treat them. Right now, frontier mathematics is next, with new discoveries coming at an impressive pace.
Imagination, imagination, imagination
The more I work with AI, the more I see that the main limit on its use is our imagination. Even if we stopped improving the technology today, there would still be far more ways to use it than we’ve thought of so far.
Part of that imaginative work is up to us. But that night with my eight-year-old, making up that story, reminded me that part of it belongs to the next generation too. As the poet Louise Glück wrote, “We look at the world once in childhood, the rest is memory.” With that in mind, we need the next generation—with their sense of adventure, open-mindedness, and lack of cynicism—to help us think wisely and freely about what we can do with these incredible technologies.
What Should My Children Do? by Daniel Susskind is published by Penguin at £20. To support the Guardian, order your copy at guardianbookshop.com. Delivery charges may apply.
Frequently Asked Questions
Here is a list of FAQs based on the perspective of a father and AI researcher written in a natural parentfriendly tone
General Beginner Questions
1 My kid uses ChatGPT for homework Is that cheating
It depends on how they use it If they copypaste an answer without thinking yes thats cheating But if they use it like a tutorasking it to explain a concept they dont understand or to quiz them before a testits a great learning tool The rule should be AI is for helping you think not for thinking for you
2 Im not techsavvy Do I need to be an expert to protect my kids from AI
Absolutely not You dont need to know how to code You just need to know the basics of how these tools work and more importantly you need to talk to your kids about how they are using them Your parenting instincts matter more than your technical skills
3 What is the biggest risk of AI for my children
I worry less about robots taking over and more about them losing the ability to think for themselves The biggest risk is intellectual lazinesskids outsourcing their critical thinking creativity and problemsolving to a machine because its faster That and the emotional impact of interacting with chatbots that seem human but arent
4 Is AI actually useful for everyday parenting or is it just hype
Its genuinely useful for the boring stuff I use it to plan meal rotations generate ideas for a 5yearolds birthday party write tricky emails to teachers and explain why is the sky blue in a way a toddler understands Its a massive timesaver for logistics but it will never replace a hug or a hearttoheart talk
Advanced Critical Thinking Questions
5 How do I explain to my teenager that AI can be confidently wrong
Explain that AI doesnt know facts it predicts what words usually come next Its like a very wellread parrot It doesnt have a brain to check if something is true so it will state nonsense with total confidence Teach them the trust but verify rule use AI for drafts but verify all facts with reliable sources before submitting
