The Future

The Industrial Revolution did bring prosperity in the end. The first three generations never saw it.

Whether AI makes the world better in the end isn't the real question. The real one is what happens to us in the meantime. And whether we'll still be around for the better part.

The favourite argument against any worry about AI transitions runs like this: “People thought everything would collapse during the Industrial Revolution too. In the end everyone was better off.” True. It just wasn’t true for the people who lived through it.

Roughly 90 years separated the first spinning machine from Bismarck’s social legislation. In between: starvation wages, child labour, slums, riots, cholera. Productivity rose for decades, real wages did not. Engels wrote down the consequences in Manchester in 1845. Only the fourth generation gained from the long-run prosperity. Generations one through three paid for the transition.

Every transition has its losers. We just like to pretend we won’t be among them. “It’ll settle down eventually” is the consolation offered by those for whom it already has. Who gets hit by structural breaks can be read off before the break arrives. The pattern has been the same for decades.

AI shifts the profile upward, into the very professions that were supposed to prove technology builds the middle class rather than dismantling it. That is exactly where the current generation of language models and agents takes aim. People who write, code, or review contracts today hold jobs that may be hollowed out in ten years. Maybe yours. Maybe ours.

The question is not whether AI creates prosperity in the long run. It probably does. The question is who sees it first and who never sees it at all. Last time it cost three generations before the answer was halfway bearable. The institutions that managed it didn’t fall from the sky. They were forced through after decades of pressure, by people who got nothing out of them anymore.

The hope that “we” will be spared is the most naive version of this story. In the 19th century, “we” meant the aristocracy at first, then the bourgeoisie, then the skilled working class. Each of these groups felt safe for a while. That time is something we don’t have.

We need to talk about this

What does this topic stir up in you?

In depth

What Allen calls the 'Engels pause'

The Oxford economic historian Robert C. Allen measured, in "Engels' Pause" (Explorations in Economic History 46, 2009), the phase in which British output per head was already rising while real wages were not. Allen dates the stagnation roughly to 1790 through 1840. An ever larger share of national income went to the owners of capital during these years, an ever smaller share to wage labour.

Real-wage gains only set in after 1840, and it took further decades before they reached most people. Anyone who entered working life around 1790 never lived to see them. The mechanism behind it is nothing exotic. Whoever owns capital pockets the productivity gains first. Whoever has only their labour gets them later, or not at all. There is no physical law that says it can't play out the same way this time.

Who are the losers this time?

The MIT labour economist David Autor ("Work of the Past, Work of the Future", AEA Papers and Proceedings 2019) traced who loses out in the big structural breaks in the US: low formal education, mid-forties and older, living away from the growth centres, work with a high routine share. The pattern has held steady since the 1980s.

AI shifts that profile upward. The MIT economists Daron Acemoglu (Nobel Prize in Economics 2024) and Pascual Restrepo show, in a 2024 NBER working paper covering the US since 1980, that automation targets precisely the tasks that pay above-average wages. That is the measured history of automation, not an AI forecast. By the same logic AI reaches the next wage brackets up: writing, research, legal review, programming, first-pass medical diagnosis. Professions long considered the safe side.

What did we learn?

Social insurance, occupational safety, union rights, progressive taxation. All of it emerged as an answer to the Industrial Revolution. Nearly a hundred years passed between the start of factory work and Bismarck's social legislation of 1883. Another fifty before the modern welfare state.

We don't have 140 years for answers like that this time. We don't even have 14.

Sources