AI Is Just a Tool : Daniel Susskind wrote a genuinely thoughtful piece in the Guardian this week about AI and his children’s education — a Dr Seuss-style bedtime story generated in seconds, an AI-narrated podcast that fooled his whole family, a call to “teach both, test both” the way calculators eventually got folded into maths education. Much of it is sensible, practical advice. But buried inside that practicality is a framing worth pushing back on: the idea that AI represents a civilizational hinge point, that “in the 21st century, the best ideas are likely to come from capable AIs instead of the heads of clever human beings.”
That line is where a genuinely useful piece about parenting and education quietly slides into something bigger and less earned. AI is a tool. A remarkable one, worth teaching children to use well — but a tool nonetheless, and the moment we start speaking about it as a replacement for human thought rather than an extension of it, we’ve handed it a frame it hasn’t actually earned.
The Calculator Comparison Is More Instructive Than It Seems
Susskind’s own analogy — the 1982 Cockcroft Report and the arrival of the calculator — is actually the strongest argument against his bigger claim, not for it. Nobody today says “the best mathematics now comes from calculators instead of the heads of clever mathematicians.” The calculator didn’t replace mathematical thought — it removed friction from arithmetic so people could spend their thinking on harder, more interesting problems. Forty years later, we don’t credit calculators with mathematical breakthroughs. We credit the mathematicians who used them well.
The same logic should apply to AI, and mostly does, right up until Susskind’s closing pages, where the framing shifts from “a tool that removes friction” to “a source of ideas that will out-think us.” AlphaFold2 solving protein folding is a genuinely remarkable achievement — but it didn’t dream up the problem, decide it mattered, or interpret what the answer meant for treating disease. Human scientists did that. The tool was extraordinary. The framing of it as an independent mind generating “the best ideas” erases the people who pointed it at the problem in the first place.
Why the Bigger Frame Actually Backfires
Here’s the practical issue with over-framing AI: the moment you tell a child (or an adult) that “the best ideas will come from AI now, not people,” you’ve quietly given them permission to stop trying to have good ideas themselves. That’s the opposite of what Susskind actually wants — his entire piece is a plea to keep basic skills sharp, to teach discernment, to make sure people can tell “when AI is being a savant or an idiot,” as he puts it, quoting Hinton.
But you can’t hold both of those positions at once without tension. If AI is where the best ideas will come from, why would a child bother developing the judgment to evaluate those ideas critically? The “teach both, test both” model only works if both halves are genuinely valued — and treating AI as the senior partner in human thought undercuts the very discernment Susskind is trying to protect.
A Tool Doesn’t Need a Mythology
None of this is an argument against AI in classrooms, bedtime stories, or the ways Susskind describes using it with his own kids — turning times tables into a unicorn-themed game, generating a custom bedtime story tuned to a five-year-old’s reading level, building an interactive map of Dracula’s plot with a daughter struggling to place a sprawling novel in context. These are genuinely excellent, imaginative uses of the technology, and worth celebrating exactly as what they are: a powerful tool used well, by a parent who stayed involved in the process the entire time.
That’s the actual lesson here, and it doesn’t require any grand narrative about AI surpassing human thought to be true. A tool amplifies whoever’s using it. It doesn’t replace them, and it doesn’t deserve credit for ideas the person using it supplied, shaped, and decided mattered in the first place.
Keep the Practical Advice, Drop the Mythology
Susskind’s practical suggestions are genuinely worth adopting: get back to literacy and numeracy basics, since they remain the foundation everything else builds on regardless of how the future unfolds. Teach AI use directly and openly rather than banning it. Test both the AI-assisted and AI-free versions of a skill, the way maths education eventually did with calculators. Choose a field because the underlying problem interests you, not because of a job title that might not exist in its current form for much longer.
All of that stands on its own, without needing the closing claim that human minds are about to be outpaced by “capable AIs.” The tool is powerful. The people using it thoughtfully — parents crafting a better bedtime story, students learning judgment alongside assistance, scientists pointing a model at a problem worth solving — remain exactly where the real thinking has always lived.
The Bottom Line
AI genuinely can make a five-year-old’s bedtime story better, a struggling reader’s phonics book more engaging, a personal tutor’s patience infinite. None of that requires believing it’s replacing human creativity or judgment — only that it’s a remarkably capable tool in the hands of someone who still has to decide what’s worth making, what’s worth asking, and what’s worth trusting. Keep the tool. Skip the mythology.
This article is a response to Daniel Susskind’s piece in the Guardian, “What Should My Children Do?” and reflects a personal perspective on AI’s role in education and society.
