Beneath the familiar cyclical story of a cooling UK labour market runs a second, structural change that could prove more lasting: employers are beginning to rewire how, and whether, they hire — and artificial intelligence is increasingly part of the reason. Entry-level and graduate roles, long the on-ramp into professional careers, are thinning out. Adzuna data shows advertised graduate vacancies fell to a record low in 2026, and a major employer survey found more than a third of firms have cut junior roles, with many pointing to automation. This matters because if AI is reshaping the bottom rungs of the career ladder, it changes not only near-term hiring but the pipeline of skills the economy will rely on for years. The hard part is separating genuine technological change from the ordinary effects of a soft economy.

The evidence that something is shifting

The most striking figures concern early-career hiring. According to the jobs site Adzuna, UK graduate vacancies fell to 8,383 in July 2026, down from 15,397 a year earlier — a drop of around 46% and the lowest level since the firm began tracking the data in 2016. Competition intensified accordingly, with more jobseekers chasing each opening.

A survey published in August 2026 by the Work Foundation at Lancaster University, conducted by Survation among 1,001 senior Business leaders in May, sharpened the picture. It found that 36% of employers had reduced entry-level roles for 16-to-24-year-olds over the previous year, and that 43% of those making cuts attributed them at least partly to Investment in AI and automation. Notably, larger firms were far more likely to cite AI than small ones. Nearly three-quarters of leaders said the labour market had become harder for young people than it was five years ago.

These are employer perceptions and survey responses, not a controlled measurement of AI's causal impact — a distinction the researchers themselves stress. But the consistency across independent sources lends the trend credibility.

Which roles and sectors are most exposed

The pattern is not uniform. The roles most affected tend to be those built around routine, structured cognitive tasks — the kind of work that generative AI and workflow automation now perform passably well. That includes elements of administration, customer service, basic data processing, entry-level coding and some junior professional-services work in areas such as research, drafting and analysis.

White-collar sectors with large graduate intakes — technology, finance, professional services and parts of media — are where the anxiety is most concentrated, precisely because their entry-level work often involves exactly the tasks AI tools handle. By contrast, jobs that depend on physical presence, manual dexterity or intensive human interaction — much of healthcare, skilled trades, hospitality and social care — are so far less directly displaced by software, even where those sectors face their own hiring pressures for different reasons.

Recruitment is itself being automated from both ends. The same Work Foundation survey found a substantial share of employers now use AI or automated systems to screen applications, changing how candidates are filtered before a human is involved.

Cyclical weakness or structural change?

The central analytical challenge is disentangling AI's effect from the state of the economy — and the honest answer is that both are at work, making attribution genuinely difficult.

The cyclical case is strong. UK payrolled employment has been falling, vacancies have drifted down for the better part of three years, and hiring costs jumped after employer National Insurance contributions rose from 13.8% to 15% in April 2025 and the National Living Wage increased again in April 2026. Faced with higher costs and soft demand, firms would be trimming junior hiring regardless of AI, because entry-level staff require Training and supervision before they add net value. Some of what is attributed to automation may simply be a downturn wearing a fashionable label.

Yet there are reasons to think a structural shift is also under way. The fall in graduate vacancies has been steeper than the broad decline in overall vacancies, suggesting entry-level roles are being hit disproportionately rather than merely moving with the cycle. Employers explicitly and repeatedly name AI investment as a factor. And the timing coincides with the rapid enterprise adoption of generative AI tools since 2023. When a specific category of work shrinks faster than the whole and the firms doing the cutting point to the same cause, it is reasonable to infer that technology is contributing, even if its precise weight cannot yet be measured.

Economists caution that history offers mixed guidance. Previous waves of automation displaced particular tasks while creating new roles elsewhere, often leaving overall employment intact but changing its composition. Whether AI follows that pattern — destroying some entry-level jobs while generating new kinds of work — or proves more sweeping is, at this stage, a scenario rather than a settled conclusion.

Why the entry-level squeeze matters beyond the individual

The concentration of the effect at the start of careers carries wider consequences. Entry-level roles are how workers acquire the experience that qualifies them for more senior positions; if firms automate away the bottom rung, they may find themselves short of experienced mid-level staff later. There is a potential paradox in cutting the training grounds that produce tomorrow's specialists.

There is also a social dimension. With youth Unemployment and the number of young people not in education, employment or training already elevated, a durable contraction in entry-level opportunities could deepen a problem policymakers are watching closely. Some employers, industry bodies and universities have begun exploring apprenticeships, restructured graduate schemes and AI-literacy training as partial responses, though these remain early and uneven.

What to watch next

The key question is whether the entry-level squeeze persists once the economy strengthens. If graduate and junior hiring rebounds as growth returns, the recent falls will look mainly cyclical. If they stay depressed even during recovery, that would be stronger evidence of a structural break driven by technology. For now, the responsible reading is that AI and automation are beginning to change UK hiring at the margins — visibly in early-career roles — while cyclical weakness amplifies and partly disguises the effect. Both forces are real; the balance between them is what the coming data will reveal.