Will AI destroy jobs or raise wages? The real economic story

Will AI wipe out jobs and crush wages, or spark a new wave of prosperity? The honest answer is: it’s not automatic either way. AI clearly makes many workers more productive, but history shows that higher productivity doesn’t always translate into better pay or better lives.

To understand what AI might do to wages and inequality, it helps to look at how economists think about technology, what actually happened during the Industrial Revolution, and why power and ownership matter just as much as the tech itself.

Two Competing Stories About AI and Wages

There are two big, conflicting narratives about AI’s impact on workers.

The intuitive fear: AI replaces workers and pushes wages down

The first story is the one most people feel in their gut. AI tools can now draft emails, summarize documents, write code, generate images, and even do basic research. That means many tasks that used to require junior staff, assistants, or entry-level hires can now be done faster and cheaper by software.

If companies can get the same work done with fewer people, they’ll hire less. That creates more competition for the remaining jobs, especially among younger or less experienced workers. More people chasing fewer roles usually means lower bargaining power and downward pressure on wages.

We’re already seeing early signs of this: fewer junior roles, more “do more with less” expectations, and AI being used as a rough but fast replacement for low-level work.

The economic theory: higher productivity should raise wages

The second story comes from standard economic theory. In most economics courses, you’re taught that wages are determined by something called “marginal productivity” – the extra value a worker adds to a business.

In simple terms: if one more worker can produce £100 worth of biscuits a day, a competitive market should push their wage toward £100. If a new technology (like AI) lets that same worker now produce £150 worth of biscuits, then in theory, other firms will try to poach them by offering higher pay. Over time, wages should rise with productivity.

From this perspective, AI is just another productivity-boosting tool, like the spreadsheet or the assembly line. If workers can do more valuable work in less time, the logic goes, their wages should increase in the long run.

So which story is right? To answer that, we need to look at history.

What the Industrial Revolution Really Tells Us

When economists think about technology and living standards, they almost always go back to the Industrial Revolution. It was the biggest technological shift in human history: the rise of factories, mechanized production, and mass manufacturing.

In the popular version of the story, new machines displaced skilled artisans, some people complained (like the Luddites who smashed machinery), but in the end everyone became richer. The lesson is often simplified to: “Don’t fear technology. It hurts in the short term, but it’s good for everyone in the long run.”

The real story is messier – and much more important for understanding AI.

Massive productivity gains, but not for workers

The Industrial Revolution massively increased how much a small group of workers could produce. One factory in Victorian East London, for example, produced around 300 million matches a day. Similar leaps happened in textiles, steel, and many other industries.

Yet for ordinary workers, especially in the early and middle phases, life was often brutal. Former farmers and rural workers were pushed off common land by enclosure laws, lost access to livelihoods, and flooded into cities desperate for work. They ended up in overcrowded, polluted urban areas, working 14–16 hour days in unsafe factories for very low pay.

Diseases were rampant, workplace injuries and deaths were common, and basic housing and sanitation were terrible. Even into the early 20th century, extreme poverty and preventable diseases like tuberculosis were still widespread among working-class families in industrial cities.

In other words: productivity exploded, but for a very long time, living standards for the majority did not.

So where did all the extra wealth go?

If factories could suddenly produce 50 times more goods than before, but workers were still poor, someone else must have been capturing the gains. That “someone” was a relatively small class of industrialists, landowners, and colonial elites.

British factories didn’t just outcompete local artisans – they undercut producers across the world. Backed by empire and military power, Britain forced other countries into trade relationships where local industries were destroyed and replaced by cheap imports from British factories.

The result was a small elite becoming unimaginably rich, while workers at home and abroad remained in poverty. In places like India and Ireland, this process was tied to mass unemployment, famine, and long-term underdevelopment. Even today, many countries that were on the losing side of that transformation are still struggling with the legacy.

This is crucial: technology alone did not lift everyone up. It made the economic pie bigger, but who got what slice depended on power, ownership, and politics.

When did living standards actually improve?

Broad improvements in working-class living standards didn’t come automatically from machines. They came from organized pressure.

As workers were concentrated into factories and industrial cities, they gained a new kind of leverage: the ability to organize, strike, and shut down production. Over many decades, this led to the rise of trade unions, labor movements, and eventually political parties that pushed for:

• Higher wages and shorter working hours
• Workplace safety laws
• Public education and healthcare
• Progressive taxation and social security

The real turning point for living standards in many rich countries came after World War II, when inequality fell sharply and wealth was redistributed through higher taxes on the rich, strong unions, and expanded welfare states.

So the lesson from the Industrial Revolution isn’t “technology automatically makes everyone richer.” It’s “technology creates the potential for higher living standards – but workers only benefit when they organize and win a fair share.”

Why the Classic Economic Story Breaks Down

Let’s go back to that neat economic theory: higher productivity should mean higher wages because other firms will compete to hire more productive workers. Why didn’t that automatically happen during the Industrial Revolution? And will it happen with AI?

You can’t expand production without customers

The theory assumes that if workers become more productive, businesses can just hire more of them, produce more, and sell more. But that only works if there’s a big pool of customers with money to spend.

Historically, rapid industrial expansion worked in two big waves because there was someone to sell to:

• First wave (British/European industrialization): Factories in Europe sold to the whole world, often by force through colonialism. European elites captured huge amounts of global wealth.
• Second wave (China’s industrial rise): Chinese factories sold cheap goods to a large, relatively wealthy Western middle class. Western consumers’ spending powered China’s growth.

In both cases, there was a large external market that could absorb a huge increase in production. That made it profitable to build more factories, hire more workers, and expand output.

Today, with AI, we don’t obviously have a similar untapped customer base. Most economies are growing slowly, governments are fiscally constrained, and both working and middle classes are financially stretched. The main group with serious spare cash is a very small, very wealthy elite – and they are unlikely to massively ramp up their spending just because AI exists.

Without a big new customer base, there’s less reason for companies to endlessly expand production and hiring, even if AI makes each worker more productive. Instead, they may simply use AI to cut costs, reduce headcount, and increase profits.

Why we didn’t get 15-hour workweeks

Back in the early 20th century, economist John Maynard Keynes famously predicted that by now, his grandchildren would be working 15-hour weeks. He saw how fast technology was improving and assumed society would choose to take some of those gains as leisure rather than endless production.

He was right that technology made it possible. But he was wrong about what we’d choose – or, more accurately, what the system would allow. In many countries, people now work as much or more than previous generations, especially when you factor in dual-income households, unpaid care work, and the rising cost of housing, education, and healthcare.

Why? Because in a highly unequal economy, most people don’t own much. A small group owns most of the assets: housing, land, energy, factories, data centers, and increasingly, the AI infrastructure itself. Everyone else has to pay them for access to basic needs – through rent, mortgages, energy bills, debt repayments, and taxes on governments that also owe money to the wealthy.

In that kind of system, work isn’t just about “choosing” how much income you want. It’s how you service a permanent obligation to asset owners. If you stop working, you can’t pay rent or your mortgage. You can’t buy food. You can’t cover your bills.

So when new technology appears, the average person can’t simply say, “Great, I’ll work less and enjoy more free time.” The gains from productivity often flow upward, while the pressure to work stays the same or even increases.

AI, Inequality, and Who Owns the Future

This brings us back to AI. Will it raise wages or lower them? Will it free people from drudgery or tighten the grip of economic insecurity?

The key question isn’t just how powerful AI becomes. It’s who owns it, who controls it, and how the gains are shared.

In a relatively equal society, where people broadly share ownership of productive assets, better technology can genuinely give individuals a choice: work less for the same income, or work the same and earn more. In a highly unequal society, where a small group owns the tools and everyone else rents access to life, new technology tends to strengthen the position of those at the top.

We’re already seeing early versions of this dynamic in AI. Large models are expensive to train and run, and are mostly controlled by a handful of big tech companies and well-funded labs. That concentration of power shapes everything from pricing to access to how AI is used in workplaces – whether as a tool to empower workers, or as a way to monitor, deskill, and replace them.

Some of the same tensions show up in other AI debates too – for example, whether advanced systems will be used to augment professionals or to cut corners and cheat, as explored in this look at Anthropic’s Mythos and AI that learns to cheat, or whether AI will truly replace skilled engineers or just change how they work, as discussed in why AI won’t replace real software engineering.

So, Will AI Destroy the Economy – or Transform It?

AI will almost certainly change the economy. It will automate some tasks, create new ones, and make many workers more productive. But whether that leads to higher wages and better lives, or more insecurity and inequality, is not a law of nature. It’s a political and social choice.

History suggests three hard truths:

• Technology alone does not guarantee higher living standards for most people.
• Without organized pressure – unions, worker movements, political action – the gains from new technology tend to flow to those who already own the assets.
• When workers do organize and win a fairer share, technological progress can support shorter hours, better pay, and higher quality of life.

So, will AI drive wages up or down? It depends less on the algorithms and more on what we build around them: labor protections, bargaining power, ownership models, tax systems, and democratic oversight.

AI has the capacity to either deepen inequality or help create a fairer, more prosperous society. The difference will come down to whether ordinary people, workers, and citizens claim a meaningful stake in the AI-driven economy – or leave it entirely in the hands of a small, already powerful elite.