Skills England 2026: AI Now Touches 70% of UK Jobs — Here Is Who Feels It First

Seven in ten UK workers now hold a job containing tasks that AI can already perform or speed up. That is not a prediction — it is the UK government's own measurement, published on 6 July 2026. Skills England's first-ever Annual Skills Report puts a hard number on a shift millions have felt but few could name: 70% of the workforce is exposed [Fact]. If you have wondered whether your role is on that list, this report has an uncomfortable answer — and, quietly, a few reasons for hope.

The number that frames everything: 70%

Skills England, drawing on the Department for Science, Innovation and Technology's 2026 analysis, estimates that 70% of UK workers are in occupations "containing tasks that AI could potentially perform or enhance" [Fact]. Read that carefully: tasks, not whole jobs. Exposure is not replacement. But exposure is where change starts, and the report is blunt that the highest exposure sits with professional, analytical and higher-paid roles built around cognitive, clerical and data-driven work — not the manual jobs automation anxiety usually targets.

There is a striking counter-signal buried in the same pages. Workers with AI skills already earn a 56% wage premium globally over peers without them [Fact], and UK employer requirements for AI roles are evolving 66% faster than for other jobs [Fact]. The message is not "AI destroys work." It is "AI is re-pricing which skills matter, and doing it fast."

Graduates and juniors are feeling it first

If exposure sounds abstract, the entry-level data is not. Online job adverts aimed at graduates ran 45% lower in 2025 than the year before [Fact]. Entry-level adverts overall fell 25% [Fact]. And the firms most exposed to AI cut employment by 4.5% overall — but by 5.8% specifically in junior positions [Fact] — while being 16.3 percentage points less likely to post vacancies at all [Fact].

Skills England is careful here, and so are we. The report itself warns it is "too early to say" whether AI is causing the graduate slowdown or merely coinciding with a weak market [Claim]. Youth unemployment hit 16.2% in the three months to March 2026 — the highest since 2015 [Fact] — with 957,000 young people not in education, employment or training. Some of that is the wider economy, not the algorithm. But the pattern of the ladder's bottom rung thinning first should worry anyone early in their career, especially in software development and computer programming, where junior tasks are exactly what today's tools handle well.

Who is exposed — and who is not

The report draws a clear line. Highest exposure: computer coding, cybersecurity, research, and the professional and analytical occupations clustered in the government's Industrial Strategy sectors. Lowest exposure: construction and hospitality — work "centred on physical activity or human interaction" that AI still struggles to touch.

Digital roles occupy a strange middle. They are simultaneously the largest growth area in Skills England's needs assessment and the fastest-transforming in day-to-day practice [Fact]. Whether that demand gets met by hiring more systems analysts and data scientists or simply by squeezing more productivity out of existing digital workers is, the report admits, genuinely unclear [Claim]. Geography compounds the divide: London, the South-East and the East of England host around three-quarters of AI company headquarters [Fact], though the West Midlands, North-West, East Midlands, Wales, and Yorkshire and the Humber have all more than doubled their AI firm counts since 2022. Skills England flags an equity risk too — without deliberate effort, existing disparities by sector, qualification level and region "may widen," concentrating AI's gains among those who already have the most access to training [Claim].

The jobs that will still grow

Here is the hope. Skills England's "priority occupations" — 22.9% of the workforce, roughly 7.6 million people in 2025 — are projected to add 1.8 million jobs by 2035, growth of around a quarter [Estimate]. The biggest increases go to care workers and home carers, programmers and software developers (yes, the very roles flagged as AI-exposed), and IT business analysts, architects and systems designers.

That apparent contradiction — software roles both exposed and growing — is the whole story in miniature. AI changes what the job is; it does not necessarily delete it. Roughly two-thirds of new entrants to these priority occupations are expected to need a Level 4 qualification or higher [Fact], which is a polite way of saying the growth is real but the entry bar is rising. Historically, more than two-thirds of learners entering these roles came through higher education, so the very pathways into them are being tested [Fact]. Construction, clean energy and adult social care are all pencilled in for continued expansion — a reminder that the safest ground is often the most physical.

The real bottleneck is skills, not robots

Strip away the headlines and Skills England's actual worry is mundane and serious: "substantial skills gaps and persistent barriers to upskilling." Most workers, it argues, do not need to become engineers — they need practical AI literacy, the ability to use, verify and safely integrate AI tools [Claim]. A smaller cohort needs deep technical skills. Employers, meanwhile, increasingly prize judgement, problem-solving, collaboration and — notably — human skills like empathy and resilience that no model replicates.

The barriers are stubborn. More than a quarter of job vacancies are already hard to fill because of skills shortages [Fact]. Business training spend fell from £49.4 billion to £44.8 billion between 2019 and 2024 — a 9% drop, or about £250 less per employee [Fact] — precisely when reskilling matters most. Small firms cite cost, time and a lack of trusted training options. The government's response includes a target to upskill 10 million workers by 2030, a £187 million 'TechFirst' package, and the AI Skills Boost programme launched in January 2026 [Fact]. Whether that closes a gap this wide is the open question the report leaves hanging.

What this means for your career

If you take one thing from Skills England's report, make it this: exposure is not destiny, but standing still is a choice. The workers pulling ahead are not the ones AI cannot touch — they are the ones who learned to use it. Practical AI literacy is now closer to spreadsheet skills than to a specialism: expected, not exceptional.

Check where your own occupation actually sits before you assume anything. The detail — task-level exposure, growth outlook and automation risk — lives on each role's page, from management analysts to personal care aides. The report's uncomfortable truth is that the bottom rung of the ladder is thinning. Its hopeful truth is that the rungs above are still being built — and there is still time to climb.

Sources

• Skills England, "Annual Skills Report 2026," published 6 July 2026 (GOV.UK). Figures on AI exposure (70%), the AI wage premium (56%), graduate and entry-level advert declines, AI-exposed firm employment effects, priority-occupation projections to 2035, skills shortages and training expenditure are drawn directly from this report and the Department for Science, Innovation and Technology (DSIT) 2026 analysis it cites.

This analysis was produced with AI assistance and reviewed by a human editor. All statistics are attributed to the Skills England Annual Skills Report 2026; the interpretation and framing are our own. The figures describe tasks and exposure, not guaranteed job outcomes.