The Jobs We Were Supposed to Lose

Every few years, a new technological frontier arrives wrapped in predictions of catastrophe. Artificial intelligence is merely the latest candidate in a long line of inventions that were supposed to sweep humanity aside. The narrative is familiar, machines will take our jobs, the labour market will collapse, and society will struggle to absorb millions of displaced workers. But if one steps away from the headlines and looks instead at the actual behaviour of the labour market, the picture that emerges is far more ordinary, and far less dramatic than the anxious narrative suggests.

Unemployment has risen modestly, but it remains well within historical lows. Job creation has slowed, but it has not reversed. The labour market is not suffering the violent shock one would expect if vast portions of human activity were being swallowed by software. The data shows mild easing, not technological devastation. If we were truly in the midst of an automation-led upheaval, the evidence would be unmistakable. Instead, the numbers reveal something far simpler, a labour market adjusting to demographic contraction and the end of an overstimulated economic cycle.

A Labour Market Without the Revolution

Those who claim AI is erasing employment imagine a process that leaves strong statistical fingerprints. If machines were suddenly doing work that millions of people once performed, we would see a rapid and unmistakable rise in productivity, because output would increase even as human input fell. We would also see a steep and persistent drop in employment, because displaced workers would be unable to transition quickly into new roles.

None of this is happening.

Productivity is improving only modestly, and employment remains resilient. Job gains continue, even if at a pace slower than during the boom years. The supposed “AI shock” is nowhere to be found. Instead, the labour market’s behaviour reflects long-standing demographic constraints and predictable cyclical cooling.

The real bottleneck is not too much automation, but too few people. With immigration lower than in previous decades and population growth slowing, the economy no longer requires the enormous monthly job gains that were once considered healthy. Employment expands at 30,000 to 50,000 jobs per month because that is all the labour supply allows, not because machines have overtaken human labour.

This shift is structural, not technological.

Machines Can Work, but They Cannot Live

What the numbers quietly affirm is a truth that economists have understood since the first machines appeared in factories, technology replaces tasks, not human economic roles. A machine can perform a calculation faster than a clerk, but it cannot decide what the calculation means. An algorithm can sort information, but it cannot determine which information matters for a firm, a household, or a society. AI, in all its sophistication, remains a tool, one that expands capability rather than eliminates purpose.

And beyond all this lies an even more elementary truth, machines do not consume. They do not eat, rest, aspire, choose, or desire. They do not buy homes, travel, raise children, or fill shops. An economy is not sustained by the production of goods, but by the human appetite for them. As long as consumption remains human, work will remain human. What changes is not the existence of work, but its form.

If AI were truly eliminating the need for labour on a large scale, we would see consumption collapse and savings surge. Instead, the opposite is true. Households continue to spend, businesses continue to hire, and the entire system continues to orbit human needs rather than machine output.

A Slowdown Misread as a Threat

Why, then, does the fear persist? Because a cooling labour market, even when caused by normal demographic and cyclical forces, creates the appearance of technological displacement. When job growth declines from 150,000 per month to 30,000, it is easy to mistake structural equilibrium for collapse. But reality is far less dramatic. The working-age population is barely growing. Immigration has tightened. Retirements continue to rise. In such a world, the labour market needs fewer new jobs to remain stable.

The calm figures are interpreted as ominous simply because expectations were inflated by years of artificial stimulus. Monetary expansion, fiscal transfers, and post-pandemic distortions created a temporary surge in demand and employment. As those forces fade, the economy returns to its natural pace. There is no evidence of mass obsolescence, only the fading echoes of a boom.

The Fear That Lives Without Proof

The narrative that AI is taking our jobs persists precisely because it offers a simple explanation for a complex moment. It allows commentators to transform a nuanced economic shift into a tidy morality tale. But the truth is not found in the rhetoric, it is found in the data. And the data shows a labour market that is not collapsing, but normalizing.

It also shows that technology is expanding capacity, not erasing purpose. Each new wave of innovation, steam power, electrification, computing, was met with the same dread. Each time, human labour did not disappear, it reorganized. AI is no different. It changes the texture of work, not the need for it.

The real danger is not that AI destroys employment, but that people and policymakers misinterpret normal adjustments as existential threats. Fear is fertile ground for poor decisions, and poor decisions can damage the labour market far more effectively than any algorithm.

The Truth Hidden in the Numbers

In the end, the labour market is telling a story entirely opposed to the one dominating public imagination. The slight rise in unemployment is not technological displacement but demographic gravity. The slower pace of job creation is not a warning of mass obsolescence but a return to equilibrium. Productivity remains steady. Employment remains broad. Consumption remains human.

The narrative may be dramatic, but the numbers speak calmly. And they say that AI is not taking our jobs. It is becoming another tool in an economy that still depends on human choice, human desire, and human purpose.

The labour market is not in the grip of a machine revolution. It is in the grip of a demographic reality, and a cultural misunderstanding.

And if we insist on misreading it, the greatest risk we face will not come from algorithms, but from ourselves.