One of the richest men on the planet has decreed that in a world drenched in artificial intelligence (AI), children need not learn math. Skip addition, subtraction, multiplication, division. Forget all of it. If today’s AI bubble bursts, he may lose his perch among the world’s most powerful individuals and the world may finally laugh at the sheer nerve of that claim. But right now, money and power—which in his case arrive together—turn every sentence he utters into scripture for the AI faithful, however wrong or self-serving.
I have heard a milder version of the same argument, offered in response to a real and urgent crisis: that AI is stunting children’s capacity to think, and that widespread use may stop that capacity from developing at all. Schools and universities everywhere feel this seismic wave. Yet, the technologists (I use it like ‘communists’) wave it away with a stock line: what about calculators? Didn’t we survive those?
It is remarkable that people who worship technology miss something this basic. Children learn numbers first. They grasp what a quantity means and what addition does long before anyone hands them a calculator. Only then does the machine become useful—as a shortcut for something already understood.
Now picture a child who has never grasped what a number is being given a calculator on day one. The digits on the screen will mean nothing to her. Worse, she’ll have no notion of what is to be done. Most of us cannot even picture that void, moving through the world with no exact concept of quantity.
That is exactly what AI threatens to do. It does not wait for understanding to form before offering to do the thinking for us. Hand a child or any student a machine that performs the very mental work education exists to build, and that capacity for thought never develops at all. We are outsourcing cognition to AI at precisely a stage when the whole purpose of schooling is to build it.
A child of privilege may survive this experiment. She has educated parents, tutors, siblings and a house full of books to compensate for machines short-cutting her thinking, though even that is no assurance.
Sadly, most have no such safety net. Hand her a chatbot that thinks on her behalf, and there is nothing to stop its insidious use. Inequality was already unfair. Dressing up this bargain as progress and handing it first to those who can least afford it makes it unfair on purpose.
If we do not stop this now, AI will not need to invade our world to conquer it. It will need no army, no Skynet moment. We will hand our world over freely, one outsourced thought at a time.
Ideology and incentive make a lethal combination. Bind them together and you get totalitarian regimes, or closer home, the wreckage social media has made of public life—big money’s incentive fused with the technologist’s ideology. Incentives, remember, are never merely economic. They are cultural, psychological and social too.
The billionaire prophet of AI fits this pattern exactly. He leads the technologist vanguard, and his fortune swells with every dollar poured into the industry. Common sense should tell us to dismiss his prophecy or simply laugh at it. But men like him who act so transparently in their self-interest could also damage the world with their ideas and proposals.
Not every AI leader speaks this way, though. We hear some of them say the opposite: we do not fully know what happens next, so we must be very careful with AI.
Far more dangerous are the pronouncements of another set of oracles: economists. Too many of them insist with great solemnity that fears of AI-driven job losses are wildly overblown. This, even as the tech industry itself and nearly every other industry sounds an alarm.
Why do economists say this? Ideology mixed with incentive, again. Their ideology is methodological rigour; their incentive is the approval of their peers, which they crave and fear in equal measure.
So, they repeat: the data shows no job losses yet. That is the economist of 1918 insisting that the data shows no effect of air transport on the movement of goods and people. After a historical rupture, past data cannot see the future. Some things simply have not happened yet, but are no less likely for it.
Many of these economists concede that AI differs fundamentally from earlier technologies, as its ability to swallow cognitive tasks rises by the day. And they still call the fear overblown because job losses have not yet shown up in the statistics. Their method cannot capture how deeply tasks and roles are embedded within organizations or how long unwinding them takes. But unwind they will. Capital wants labour gone, and AI will let markets deliver exactly that, one restructured company at a time.
Cornered, these economists retreat to a second line: scary job-loss forecasts are just valuation theatre, dreamt up to justify the AI industry’s absurd valuations. What they refuse to admit is that both things can be true at once: a bubble in AI company valuations and a coming tsunami of real job losses. One does not cancel the other.
Push further and you get the last resort: every new technology has destroyed some jobs and created others. But not one of them can name what jobs might replace even a fraction of what will be lost.
Between the technologists’ sales pitch and the economists’ methodological piety, both sets of oracles are gambling with our future. Neither is willing to bet with their own.
The author is CEO of Azim Premji Foundation.