AI is boosting UK productivity, but it's coming at a cost for workers

By Mihir S Sharma

 

Is the British economy doing well or poorly? At first glance, the picture appears muddled, with data pulling in different directions.

 

On the pro side: The Office for National Statistics says that June growth, at 0.3%, was higher than expected. Business investment, long the UK economy’s weakest link, was up 1.7%. The FTSE 100 has registered gains over six successive quarters. And overall wage growth in the second quarter – driven in part by National Health Service pay raises – was also slightly higher than expected.

 

But during that same period, the private-sector portion of wage growth, excluding bonuses, increased at the slowest pace since the pandemic; the number of job vacancies, at 707,000 over the quarter, was also lower than it had been since the Covid years. From April to June, the unemployment climbed to 4.9%.

 

 

These are not inconsistent readings. In fact, they tell a coherent — if uncomfortable — story. They show that, since Labour took office in 2024, Britain has seen a K-shaped recovery: The top end of its economy is doing better than it has in years, but the bottom is struggling.

 

The conventional wisdom is that Britain faces a productivity problem. Its workers simply aren’t efficient enough, and that hampers growth. Recent work from economists at the London School of Economics — two of whom worked until recently for then-Chancellor Rachel Reeves — turns that assumption on its head. They find “an annualized productivity growth of 1.6 per cent between 2024 Q3 and 2026 Q1, compared with only 0.3 per cent over the decade preceding 2024 Q3.” And a new analysis from Morgan Stanley shows private-sector productivity gains accelerated in the past few months.

 

But that’s not entirely good news. One factor driving the LSE researchers’ results is the fact that the administrative databases they used, particularly the UK’s Pay As You Earn tax records, suggest the number of people in work declined by 133,000 over the period they analyzed, while the official statistics say it actually increased by 377,000.

 

So, there are fewer jobs around, but those in them are producing more and – particularly if they’re in the public sector – being paid more, as well.

 

Why is this happening? The study’s authors think it’s probably the first sign of what artificial intelligence will do to the economy overall, rather than any shift specific to low-wage jobs. But others have found that hiring is falling fastest for roles that were already paid the least.

 

Both things could be true. There are multiple reasons why businesses might not want to take on a relatively unskilled employee right now, and uncertainty about AI is certainly one. Economists from the Bank of England say that vacancies have dropped fastest in professions most vulnerable to AI substitution. Customer service jobs, for example, have seen online advertisements drop by an average of 23% a year since 2023.

 

Finding a job was never easy, but it’s a particularly nightmarish experience right now, for jobseekers as well as those on the other side of the interview table. Unmanageable numbers of people are applying for each role, many using AI-constructed CVs of dubious integrity. LinkedIn said that applications went up 45% last year and that it was processing 11,000 of them a minute.

 

Genuine applicants are being put through the wringer, enduring multiple interviews even for entry-level openings as hiring managers attempt to filter out fraud. The hurdles are highest for those without experience, including career changers and recent graduates (youth unemployment is higher than it has been for more than a decade). Even people employers might want to hire are put off by the process.

 

Let’s suppose that there’s both a good story — investment, fiscal stability, and AI improvements for large companies and more experienced workers — and a bad one, in which the lower end of the labor market is being inundated by AI sludge and paralyzed by uncertainty. The crucial point, as far as policymakers are concerned, should be that this is not transitory. Instead, it’s likely a transition. We are shifting to a world in which these two groups might fare very differently.

 

Both UK political wings, blindly following instincts hardened over decades, are approaching this change the wrong way. Labour, through some provisions of its Employment Rights Bill and by raising employer National Insurance contributions, made it financially harder to take a punt on an unproven 19-year-old at precisely the moment when the payoff for doing so has become harder to predict. Meanwhile, the Tories are – incredibly – still blaming European Union regulations, and Reform UK tries to shift the fault to immigrants.

 

Britain’s nascent top-end revival is worth protecting. But its policymakers need to think harder about how to get young people and those in AI-exposed professions in on those productivity gains. That might require flexible thinking and flexible regulations.

 

Everything from housing reform to small-business regulation needs to be reoriented toward this new scenario. Mobility matters as much as skills; young people need to be able to move to where jobs are, and must be able to afford to live in those places. And smaller employers, fearful of AI, might need to be given better security to experiment with more hiring, with new and lighter-touch probationary or temporary arrangements.

 

Labour’s apprenticeship reform — better-funded places for under-25s, for example — is a useful start. But there’s a lot more the party — and its opponents — could and should be doing. Without open minds and fresh ideas, we’re in real danger that AI will transform the job market before British politics even starts questioning its too-long-held assumptions.