Raising the Generation That AI Cannot Replace
- 1 hour ago
- 4 min read
Why giving students less technology might be Singapore's greatest competitive advantage
I had the opportunity to coniduct a 1 hour data workshop to a group of 14 year olds. There is a student I will not forget in a hurry.
During a data skills workshop I recently delivered at a learning festival, I asked a room of 14-year-olds whether they used AI. Every hand went up. I asked what for. One answer stopped me cold.
"For school work."
I gently pointed out that AI has been known to give wrong answers. The student's response was immediate, and completely unbothered:
"I don't care if it's right or wrong. The homework was done."
That moment crystallised something I have been quietly worried about for a while.
We might be solving the wrong problem.
Singapore's push to infuse AI into primary school education is well-intentioned. In a world increasingly shaped by intelligent machines, the instinct to prepare the next generation early makes intuitive sense. But there is a growing body of evidence suggesting we may be building a generation that is fluent in using AI — but wholly unable to question it.
And that distinction matters enormously.
What the research is telling us
The evidence has been building for some time. A widely cited 2014 study by Mueller and Oppenheimer found that students who took notes by hand significantly outperformed those who typed — not because they wrote more, but because the slower pace of writing forced them to process and paraphrase ideas in their own words. The struggle was the learning.
More recently, MIT researchers found that heavy reliance on AI-assisted writing was associated with reduced brain engagement — and over time, diminished capacity for independent thought. The brain, like a muscle, responds to the demands placed upon it. Remove the demand, and atrophy follows.
Meanwhile, Sweden — once a global pioneer in classroom digitalisation — has reversed course. After observing declining literacy and reading comprehension scores, Sweden's National Agency for Education began recommending a return to printed textbooks and handwriting. The experiment had run its course, and the data was unambiguous.
The broken loop no one is talking about
There is a structural problem at the heart of the AI-in-education agenda that rarely gets examined.
The modern workplace increasingly needs workers who can validate, direct, and critically evaluate AI outputs. AI produces answers at speed, but it also hallucinates, misinterprets context, and reflects the biases of the data it was trained on. A human in the loop who can recognise a flawed output — and know why it is flawed — is not optional. It is the last line of defence.
But here is the uncomfortable truth: if the next generation learns to think with AI before they learn to think without it, they will not develop the foundational reasoning skills to serve that function. They will press "generate" and trust the output — exactly as my student did.
This is not a pessimistic view of young people. It is a predictable consequence of removing productive struggle from the learning process. We learn to validate answers by first learning how to find them. You cannot audit a process you have never performed yourself.
The case for going back to basics
This is where I will say something that may be unpopular in an era of smart boards and AI tutors:
There is almost nothing in a primary school curriculum that genuinely requires AI. And very little that truly requires technology at all.
Mathematics builds the reasoning framework. Language builds the capacity to structure and communicate thought. Science — conducted through physical experiments, not digital simulations — builds the intuition to observe, hypothesise and be surprised by reality. History and literature build the empathy to understand what it means to be human in a complex world.
These are not outdated skills. They are the very skills that make a person capable of working with AI rather than being replaced by it.
Weaning a device-dependent generation off screens for six to eight hours of their school day is not deprivation. It is arguably the most future-ready thing we could do.
The competitive advantage hiding in plain sight
Here is the strategic possibility that Singapore, of all places, is well-positioned to see:
The world is producing graduates who are fast, digitally fluent, and deeply dependent on AI to think for them. If Singapore were to deliberately produce graduates who are slower to reach for AI — but far more capable of reasoning independently when they do — that cohort would be extraordinarily valuable.
Not in spite of their restraint. Because of it.
The organisations navigating the AI transition most successfully will not be those with the most AI tools. They will be those with the most humans who can think clearly about what AI is telling them. Critical thinking, contextual judgement, and the ability to ask better questions — these are the skills in shortest supply.
AI is also, by design, user-friendly. And it changes rapidly. There is a reasonable argument that the optimal time to learn to use AI is when a student has already developed the cognitive foundation to engage with it critically — not before.
What I took away from one hour with 14-year-olds
Teaching data skills to teenagers and teaching working professionals are, I discovered, not as different as you might expect. The concepts are the same. The "aha" moments look the same. The resistance looks the same too.
What struck me most was this: when students were given a problem that mattered to them — their own school's tuck shop data — and the tools to investigate it themselves, they were engaged. Curious. Even competitive. No AI needed.
The spark was lit not by giving them answers, but by giving them a reason to find the answers themselves.
That is what education is for. And it is something no algorithm, however sophisticated, can replicate.
The greatest gift we can give the next generation is not access to the most powerful tools. It is the confidence and capability to think without them — so that when they do reach for those tools, they remain firmly in charge.
Comments