The mechanisation of textile mills, the introduction of electricity, the rise of computers and the expansion of the internet were all accompanied by predictions of widespread unemployment. Some jobs disappeared. Many changed. Entirely new industries emerged. Yet employment itself proved remarkably resilient.

Artificial intelligence has revived the debate with fresh urgency. Headlines routinely ask whether AI will replace accountants, lawyers, consultants, marketers and software developers. Businesses are experimenting with AI-powered automation at unprecedented speed. Government policy appears to be betting on adaptation rather than protection. More than £3 billion is being invested in AI capability, infrastructure and workforce skills, despite only a small minority of AI-using businesses reporting reductions in headcount.

But there is a possibility that deserves greater scrutiny. What if we have misdiagnosed the problem entirely? The emerging evidence suggests the UK’s biggest employment challenge may not be AI itself. It may be a labour market that is struggling to adapt to technological change.

The Wrong Question

Most discussions begin with a simple assumption: ‘AI can perform task X, therefore job X is at risk’.

It is an understandable conclusion. Today’s AI systems can write reports, analyse data, summarise research, generate software code and perform many tasks previously associated with knowledge workers.

Yet there is a crucial distinction between capability and deployment.

Anthropic’s March 2026 US labour market research found that real-world AI usage remains significantly below what current models are theoretically capable of achieving. In other words, AI can do far more than organisations are currently asking it to do. The gap between technical capability and practical adoption remains substantial.

History suggests this should not surprise us. Electricity existed for decades before factories were redesigned around it. The internet was available long before businesses transformed their operating models. Cloud computing took years to move from technical possibility to business necessity.

Technology rarely changes the economy overnight. Organisations do.

Britain’s Productivity Problem Predates AI

Long before ChatGPT, Anthropic, etc. entered the public consciousness, the UK faced a persistent productivity challenge.

Since the financial crisis, productivity growth has lagged behind many comparable economies. Businesses continue to wrestle with rising costs, skills shortages, demographic pressures and intense international competition.

Against that backdrop, AI is often presented as a solution rather than a threat.

This distinction matters. If a company introduces AI to reduce repetitive administration, accelerate decision-making or improve operational efficiency, the primary objective is productivity improvement. Any employment effects are seen as secondary consequences.

Seen through this lens, AI may be less a cause of labour market disruption than a response to deeper economic pressures.

The Curious Case of the Missing Talent

At the same time, another contradiction is emerging. Business leaders frequently report difficulties recruiting skilled workers. Sector leaders continue to cite shortages in digital skills, engineering capability and leadership talent.

Yet many of the same organisations are reducing entry-level recruitment. Research increasingly suggests that AI-exposed sectors are hiring fewer junior workers while continuing to seek experienced professionals.

This creates a question that should concern every boardroom: How can organisations complain about a shortage of experienced talent while simultaneously reducing the intake of future talent?

The issue is not new. Following previous economic shocks, many employers reduced graduate recruitment in the pursuit of short-term efficiency. Years later, those same organisations often found themselves facing capability gaps, leadership shortages and succession challenges.

Today’s AI debate risks repeating the pattern. The concern is not simply that AI might eliminate jobs. The concern is that businesses may be weakening the pipeline through which future expertise is developed.

The Experience Paradox

This may be the most important workforce issue created by AI. Traditionally, junior employees performed work that more experienced professionals no longer had time to do. Researchers gathered information. Junior lawyers reviewed documents. Graduate accountants reconciled data. Analysts prepared reports. Junior engineers documented systems and processes.

Much of this work was repetitive. Some of it was administrative in nature. None of it was glamorous. But it served a critical purpose. It was how expertise was built. The profession effectively paid people to learn.

Today, many of those tasks can be performed more quickly by AI. At first glance, this appears to be an unambiguous gain in productivity. Yet a deeper question emerges: If AI performs the work through which people traditionally learned their profession, how do they gain experience?

Businesses may inadvertently be creating what could be called the Experience Paradox. The more successfully organisations automate junior work, the fewer opportunities exist for junior workers to become experienced professionals.

Over time, this risks creating a self-inflicted talent shortage. Technology may remove some of the apprenticeship layer of modern knowledge work. The challenge for leaders is deciding what replaces it. This is not fundamentally a technology problem. It is a workforce design problem.

What Government Policy Reveals

Government spending often reveals assumptions more clearly than public statements. If policymakers genuinely believed AI would trigger widespread unemployment in the immediate future, we might expect substantial investment in labour market protection schemes.

Instead, much of the UK’s investment is focused on AI adoption, workforce skills, training, research and innovation. Billions of pounds are being directed towards AI capability, infrastructure, education and workforce development.

That is a significant signal. It suggests policymakers view the greater risk as Britain failing to adopt AI rather than Britain adopting it too quickly.

Whether that judgement proves correct remains to be seen. However, it reflects a view that AI’s economic opportunity currently outweighs its labour market threat. Business leaders would be wise to consider why.

What the Evidence Does Not Yet Show

A sober assessment requires caution. There is currently limited evidence that AI will caused mass unemployment in the UK. There is limited evidence that AI has triggered a collapse in employment opportunities across the economy. There is limited evidence that AI alone explains changes in hiring patterns.

What the evidence does suggest is more nuanced. We are seeing: shifts in task allocation; reduced demand for some entry-level work; organisational restructuring; slower hiring in selected knowledge-worker roles; growing pressure on traditional career pathways.

These developments matter. But they are fundamentally different from the apocalyptic narratives that often dominate public debate. The labour market is changing. That does not necessarily mean it is collapsing.

The Bigger Question

Perhaps the most important question facing organisations is not whether AI will replace people. It is whether organisations are redesigning work thoughtfully enough to ensure people still have pathways into expertise.

Every generation of leaders inherits a workforce built by the previous generation. The experienced professionals businesses seek today were once inexperienced graduates learning their craft through routine work. If that work disappears, what replaces it? The answer will shape not only productivity, but leadership pipelines, organisational resilience and long-term competitiveness.

The future of work may depend less on what AI can do and more on how organisations choose to adapt around it. The real challenge may not be technological disruption. It may be whether businesses can continue creating expertise in a world where expertise is no longer built in the same way.

Questions Every Business Leader Should Be Asking

Are we solving a productivity problem or reacting to an AI trend? Which tasks are we automating, and what learning opportunities disappear with them? If entry-level work changes significantly, how will future expertise be developed? Are we reducing graduate recruitment in ways that could create future capability gaps? Do our workforce plans assume experienced talent will always be available?

How will we redesign career pathways in an AI-assisted organisation? Are we measuring AI capability or genuine business adoption? What skills will become more valuable as routine knowledge work becomes automated? How will we maintain organisational memory, judgement and professional development? Could the biggest risk from AI be organisational rather than technological?

For many organisations, that final question may prove to be the most important of all.