The Economic Inevitability of AI

When wages get expensive, capital gets creative. Four centuries of evidence show the same feedback loop at work today.

The Pattern That Keeps Repeating

For the past four hundred years, wages and technology have chased each other in a recurring loop. When labour markets tighten and pay rises, businesses find it profitable to invent machines that replace workers.

Those machines raise output per worker but destroy the livelihoods of people whose skills they have made redundant. Average wages stagnate, inequality widens, and profits swell. Only once the old economic structure has been fully displaced does a new equilibrium form in which wages and productivity rise together again, until the next wave begins.

This is the cycle that powered the shift from cottage spinning to factory spinning in the 1770s, from artisan workshops to assembly lines in the early 1900s, and from factory floors to service counters after 1973. The names change. The logic does not.

How the Loop Works

• Expanding markets tighten the labour supply, pulling more workers into employment, driving pay upward, and raising the cost of production for employers seeking to maintain margins.

• High wages relative to the price of capital make it worthwhile for firms to develop labour-saving technology, because the payoff from eliminating an expensive worker exceeds the cost of building the machine that replaces them.

• New technology displaces incumbent producers gradually, creating a two-tier economy in which those working with the new technology earn more, while those clinging to old methods see earnings collapse. This is the K shaped economy we're seeing.

• Average wages stagnate even as everyone gets more productive, because gains in the new sector are offset by losses in the old one, and profits capture the difference.

• The old ways are eventually competed out of existence, removing the drag on average earnings, and a new equilibrium takes hold in which wages and productivity once again rise in step.

• Product innovation and falling prices sustain demand preventing saturation, ensuring the new system has room to expand, moving down the value chain and displacing those earning less lower down.

• Globalisation amplifies the upswing and the pain, opening vast new markets but also exposing domestic workers to low-wage competition from abroad.

  
  

Where AI is in that Loop Today

Artificial intelligence is the new labour-saving technology, and its trigger, as ever, is the high cost of human workers. In February 2026, Federal Reserve Governor Michael Barr described AI as increasingly likely to become a general-purpose technology on a par with the steam engine, electricity, and the personal computer. By December 2025, 17% of American businesses reported using AI, rising to 30% among firms with more than 250 employees.

The displacement is already visible. A King's College London study found that firms most exposed to AI cut total employment by 4.5% on average, with the effect concentrated in junior positions, which fell by 5.8%. Workers with AI skills now command a 56% wage premium over peers who lack such skills, up from 25% the previous year. This is the two-tier economy the historical pattern predicts.

The AI Transition

• AI is the spinning jenny of the service economy, making it cost-effective to automate tasks previously too reliant on human judgement for machines, from customer support to legal research to medical diagnostics.

• The displacement falls first on the young and the mid-skilled, just as the power loom destroyed handloom weavers; early-career workers in AI-exposed occupations are already experiencing declining employment.

• Capital is being substituted for labour at an accelerating rate, driven by falling computing prices much as falling prices in steam and iron once drove factory adoption.

• The labour share of national income continues to fall, with profits accruing to capital owners while median wages are static.

• A K-shaped labour market is forming, splitting into high-pay roles for workers who can build and direct AI systems, and low-pay, precarious work for those who cannot.

• The gig economy acts as the modern equivalent of the handloom sector, absorbing displaced workers into insecure roles that suppress average wages.

• Institutional responses lag behind the pace of disruption, with weak labour laws and declining union membership driving a wedge between productivity growth and wage growth.

  
  

What Lies Ahead

History offers both reassurance and a warning. Every previous episode of creative destruction has given way to a new equilibrium of broadly shared prosperity. But the interregnum can last a very long time.

The original pause, during which British average wages stagnated while output per worker rose, lasted roughly seventy years. The current UK wage stagnation has already run for more than fifteen years.

In the near term, AI adoption will accelerate in professional and white-collar services where labour costs are highest, mirroring the way the spinning jenny was adopted first where wages were highest in 18th-century England.

The wage premium for AI skills will widen, reinforcing income polarisation. The gig economy will grow as a default destination for displaced workers, keeping measured unemployment low while underemployment rises.

Further out, the question is whether governments, employers, and workers can rebuild institutional machinery fast enough to match the speed of technological change. The Age of Manufactures delivered rising living standards for almost a century because product innovation kept creating new demand, collective bargaining distributed the gains, and public investment in education and research supported the transition.

When those structures were weakened or absent, the result was destitution for millions even as the economy grew.

The loop does not break. It only waits for society to catch up.