Will My Job Be Replaced by AI? A Data-Driven 2026 UK Risk Outlook
Will my job be replaced by AI? Our 2026 UK outlook pairs ONS ASHE 2025 salaries with named AI exposure studies for 20+ roles — see who is safe and who is not.
Will my job be replaced by AI? The short answer for UK workers in 2026.
If you have asked yourself, "will my job be replaced by AI?" — you are not alone, and the honest answer for most UK workers is probably not in the next five years, but the work itself is going to change a lot. Generative AI is already doing parts of jobs that used to be done by humans, especially in admin, legal support, customer service, and entry-level analysis. The first wave is hitting fastest in roles where the day looks like reading a document, writing a draft, or pulling numbers from a system.
But the picture across UK careers is more uneven than the headlines suggest. The Institute for Public Policy Research (IPPR) finds that around 11% of UK tasks are already exposed to existing generative AI, rising to 59% if firms integrate the technology more deeply (Transformed by AI, March 2024). That is task-level exposure, not job replacement. Most jobs in 2031 will still exist; they will just look different.
This guide pairs the most credible UK AI-exposure research with 2025 ONS Annual Survey of Hours and Earnings (ASHE) salary data, so every risk claim is anchored to a real UK salary number. We split 20 occupations into three tiers — high, medium, and lower exposure — and explain where sources disagree so you can make your own call.
The five-year framing. When researchers say a role is "at risk," they usually mean tasks within that role are being automated now or in the next 5–10 years. That changes pay, hiring volume, and what entry-level looks like long before anyone is "replaced."
How we ranked the risk
We did not invent a new score. We blended three UK-relevant sources, plus two well-known international studies for context, and explained where they disagree.
Primary sources (UK-focused):
- IPPR (March 2024) — Transformed by AI analyses 22,000 work tasks across the UK economy. Task-based, not job-title based. Distinguishes "here and now AI" exposure from "fully integrated AI" exposure.
- UK Department for Education (November 2023) — The impact of AI on UK jobs and training. Builds an AI Occupational Exposure (AIOE) index across 365 SOC categories, identifying the most and least exposed occupations and noting that London and the South East face higher exposure because of how professional jobs cluster there.
- ONS (May 2026 release covering April 2025 data) — Annual Survey of Hours and Earnings (ASHE 2025), the gold-standard UK salary source. Median full-time gross annual pay in the UK is £39,039.
Context sources (cited carefully):
- PwC UK AI Jobs Barometer (2024) — Higher exposure estimates than IPPR; uses a different methodology.
- ILO–NASK Global Index of Occupational Exposure (2025) — Frames the shift as transformation, not replacement: roughly 1 in 4 jobs globally face significant change.
- Frey & Osborne (2013) — Influential but pre-GenAI. Helpful only as a historical anchor. Their 35% figure for UK jobs at risk uses a much broader definition of "computerisation" than today's AI debate, so we do not lead with it.
A few rules we followed. Every salary figure below comes from ONS ASHE 2025 unless we say otherwise. Where an ONS median is suppressed (too few respondents to publish safely), we say so — we never show it as £0. Where sources disagree on risk, we say so and explain why rather than pick a winner. And we use transformation more than replacement — that is the language IPPR and the ILO use, and it matches what is actually happening in most roles.
Highest-risk UK roles: where AI is already doing the work
These roles do a high share of tasks that current generative AI handles well: drafting text, pulling structured data, summarising documents, answering routine queries.
Table A — Higher-risk UK roles: 2025 median salary and AI exposure tier.
| Role | 2025 median salary | Exposure |
|---|---|---|
| Data entry clerk | £31,317 | Very high |
| Telemarketer | £22,747 | High |
| Paralegal / legal assistant | £27,746 | High |
| Customer service representative | £19,000–£26,000 | High |
| Copywriter / marketing specialist | £28,700–£45,000 | High |
| Accountant / bookkeeper | £32,117–£39,900 | High (contested) |
| Management consultant | ~£50,000+ | Very high |
| Business analyst | £45,000–£60,000 | Very high |
| Financial analyst | £42,000–£55,000 | High |
| Marketing manager | £38,000–£55,000 | Medium–high |
| Graphic designer | £22,000–£28,000 | Medium–high |
| HR officer / recruiter | £28,000–£35,000 | Medium |
| Secondary school teacher | £32,916–£45,352 | Medium |
Sources and notes. Salary figures: ONS ASHE 2025 except where noted (management consultant, business analyst, financial analyst, marketing manager, HR officer — market estimate; graphic designer — entry-to-mid range from ONS ASHE category-level data; secondary school teacher — England STRB main pay range, September 2024 awards in force). Exposure tier sources: IPPR 2024, PwC 2024, ILO 2023, DfE 2025, Frey-Osborne (2013 reference, contested for accounting).
What this tier has in common. Most of the day is reading, writing, classifying, or transforming structured information. That is the home turf of large language models in 2026.
Paralegal salaries in the UK sit at the lower end of the white-collar pay scale, and the role does a lot of work — discovery review, contract markup, statute lookup — that GenAI tools can now do in seconds. IPPR flags legal support roles as among the most affected in the "here and now" wave. Senior paralegals who own client relationships and case strategy are far less exposed than juniors doing document triage.
Accountants and bookkeepers are the classic example of sources disagreeing. Frey & Osborne (2013) put accountants at a 0.94 probability of computerisation — one of the highest in their dataset. IPPR (2024) and the DfE (2023) rank them more moderately, because the parts of the job that are exposed (reconciliation, bookkeeping, first-pass tax return drafting) are not the parts that drive senior pay (tax strategy, audit judgment, client relationships). The honest framing: routine accounting work is being absorbed by software; advisory accounting is not. For a sense of where these roles sit on the wider pay distribution, see our UK average salary 2024 breakdown.
Copywriters, market researchers, and consultants are exposed because their primary deliverables (decks, briefs, reports) are exactly what GenAI is best at first-draft producing. The work is not disappearing, but the shape of it is shifting — fewer juniors writing first drafts, more senior reviewers editing AI output. Expect entry-level hiring volumes here to fall before headline salaries do.
Mid-risk UK roles: AI changes how the work is done, not whether it gets done
Table C — Mid-risk UK roles: 2025 median salary and AI exposure tier.
| Role | 2025 median salary | Exposure tier |
|---|---|---|
| Software developer | £36,207 | Medium |
| Marketing manager | £39,236 | Medium |
| Graphic designer | £22,000–£28,000 | Medium–high |
| HR officer / recruiter | £28,000–£35,000 | Medium |
| Secondary school teacher | £32,916–£45,352 | Medium (admin + content) |
Sources and notes. Software developer and marketing manager: ONS ASHE 2025. Graphic designer: entry-to-mid range from ONS ASHE category-level data plus market data. HR officer / recruiter: market estimate. Secondary school teacher: England STRB main pay range, September 2024 awards in force. Exposure tier sources: IPPR 2024, DfE 2025, expert review.
Software developer pay in the UK comes in below the lazy stereotype — £36,207 is the ONS ASHE 2025 median for the SOC 2135 grouping, and seniors earn well above that. Coding assistants like GitHub Copilot already write boilerplate, scaffold tests, and explain unfamiliar code. The bits the AI does not do well (system design, debugging at the seams between systems, judging trade-offs against a business goal) are exactly the parts that pay senior salaries. Expect entry-level developer hiring to feel the squeeze first; senior pay is more insulated — see our deeper software engineer salary breakdown for how the pay curve flattens past mid-career.
Marketing managers are exposed in the deliverables (copy, ad creative, segmentation analysis) but protected in the work that justifies the salary (campaign strategy, budget calls, relationships with leadership). The DfE notes that office-based managerial roles concentrated in London and the South East face higher exposure than their counterparts in other UK regions — partly because of role mix, partly because of how much of the work is documents.
Secondary teachers are a useful test case. The teaching itself — classroom management, pastoral care, judging whether a 14-year-old has actually understood quadratics — is not something current AI can do. But lesson planning, marking, resource preparation, and report writing all sit in AI's sweet spot, and that is roughly 30–40% of a teacher's working week. Teacher pay in the UK is set on a national pay scale rather than by ASHE, so the figures above are England STRB main and upper range bounds. AI is more likely to give teachers their evenings back than to push them out of the profession.
Graphic designers sit higher in this tier than people expect because image generation tools (Midjourney, DALL·E, Adobe Firefly) cover a lot of production-grade asset work. The strategic, brand-defining parts of design — visual identity, complex layouts, brand guidelines — are far less exposed than asset production. Junior designers who only do production work are the most exposed group inside this role.
Lower-risk UK roles: physical work, judgement, and care
These roles do work that current AI cannot do well: they require physical presence, real-time judgement under uncertainty, and human-to-human trust.
Table B — Lower-risk UK roles: 2025 median salary and AI exposure tier.
| Role | 2025 median salary | Exposure |
|---|---|---|
| Registered nurse | £33,000–£42,000 | Low (clinical) / Medium (admin) |
| Electrician | ~£39,000 | Low |
| Plumber | £33,000–£39,000 | Low |
| Care worker | £26,000–£28,000 | Low |
| Primary school teacher | £32,916 starting | Low–medium |
| Construction supervisor | £35,000–£45,000 | Low |
| Chef | £25,000–£35,000 | Low–medium |
Sources and notes. Registered nurse: NHS Agenda for Change 2025, Bands 5–7. Electrician: ONS ASHE 2025 (close to UK median for full-time work). Plumber: market data + ONS ASHE category-level. Care worker: market data, full-time equivalent. Primary school teacher: England STRB main range minimum, 2025. Construction supervisor: market estimate. Chef: market data. Exposure tier sources: DfE 2025, expert review.
NHS nurse pay is set by the Agenda for Change pay framework rather than ASHE, so the range above reflects 2025 Band 5 (newly qualified) through Band 7 (specialist or senior practitioner). The clinical work — assessing a deteriorating patient, breaking bad news, managing risk during a complex shift — has no current AI substitute. Admin (rota balancing, charting, discharge summaries) is exposed and may genuinely free up nursing time over the next five years. For the full Agenda for Change picture, see our nurse salary guide.
Electricians, plumbers, care workers, and construction supervisors share a common feature: every site, every patient, every job is slightly different, and getting it wrong has physical consequences (a fire, a flood, a fall). Robotic dexterity is improving but is nowhere near the price–performance needed to replace skilled trades at scale by 2031. AI is far more likely to help these workers with quoting, scheduling, and compliance paperwork than to take their core job. UK pay for electricians, plumbers, and care workers is published on Salary Scout for every English region, Scotland, Wales, and Northern Ireland.
Primary teachers are listed at lower risk than secondary because so much of the work is interpersonal and physical — managing a Reception class is not a task GenAI can do remotely. The STRB main-range starting figure (£32,916 in 2025, England) is set nationally and is not derived from ASHE.
Geographic caveat from the DfE. "Lower exposure" is a UK-wide average. Some lower-risk roles are concentrated in regions where local labour markets are tighter; some higher-risk roles are concentrated in London and the South East. The risk to your specific job depends on the local mix as well as the national score.
What this means for your career and your pay
If you are reading this because you are mid-career and worried — here is the practical view, based on what the data actually shows.
1. The most exposed tasks are not the same as the most disappearing jobs. In the high-risk roles above, the entry-level rung gets thinner and senior pay starts to depend on doing the parts AI does not do well. If you are a junior paralegal, accountant, or analyst, building judgement, client trust, and domain depth is the protection — not learning to prompt better.
2. "Transformation" is the right word more often than "replacement." Both IPPR and the ILO emphasise this. Even in the highest-risk roles above, the realistic 5-year scenario is fewer hours of routine work, more AI-augmented output per worker, and a slow shift in who gets hired — not mass redundancy. The exception is roles where the entire job is one exposed task (data entry being the clearest example).
3. Pay does not always follow risk in the direction you expect. Highly-exposed roles like financial analyst, business analyst, and management consultant still pay well above the UK median because the senior end of the role is heavily protected. Low-exposure roles like care work and kitchen work sit below the UK median because labour supply, not AI risk, sets the wage. Career-risk and pay-risk are not the same thing.
4. Look at the task mix of your specific job, not just the title. A senior accountant doing tax strategy has a different risk profile to a bookkeeper at the same firm. The IPPR task-level methodology is the most honest of the bunch precisely because it does this.
5. The boring advice is the right advice. Future-proof careers in the UK are not a single list. They cluster in two places: where the human is the value (relationships, care, judgement in physical environments) and at the senior end of transforming knowledge jobs, where editing, judgement, and accountability still matter.
If you want to compare your role to the UK median (£39,039 in ASHE 2025), our salary pages cover hundreds of UK occupations with the ONS data behind them — a useful sense-check before any career conversation.
Methodology and sources
Salary data. "ONS ASHE 2025" figures come from the Office for National Statistics' Annual Survey of Hours and Earnings covering April 2025 pay, published May 2026. Where an ONS median is suppressed (too few respondents to publish safely), we either omit the figure or note the suppression — we never substitute £0 or a guess. Teacher pay uses the England School Teachers' Pay Review Body (STRB) 2024–25 main and upper ranges. NHS nurse pay uses the Agenda for Change framework, 2025 rates, Bands 5–7. Roles marked "market estimate" are drawn from market salary data where the granular SOC 4-digit ASHE cell is not consistently published, and are labelled so readers know they are not directly from ASHE.
AI exposure scores. Drawn from named, dated studies: IPPR's Transformed by AI (March 2024), the DfE's The impact of AI on UK jobs and training (November 2023), PwC's UK AI Jobs Barometer (2024), and the ILO–NASK Global Index of Occupational Exposure (2025). Frey & Osborne (2013) appears only as a historical anchor, not a 2026 forecast.
Things we deliberately do not claim. "X% of UK jobs will be lost by Y" — those percentages vary wildly by methodology and time horizon. "AI will replace teachers / nurses / lawyers" — it will change parts of those jobs, not, on current evidence, the core of them. "Frey & Osborne proved 35% of UK jobs will go" — that 2013 paper measures susceptibility to broad computerisation over a 10–20 year window before ChatGPT existed; citing it as a 2026 GenAI forecast is misleading.
Full citations.
- IPPR (2024). Transformed by AI: How generative artificial intelligence could affect work in the UK. Institute for Public Policy Research.
- UK Department for Education (2023). The impact of AI on UK jobs and training.
- Office for National Statistics (2026). Annual Survey of Hours and Earnings: 2025. Published May 2026.
- PwC UK (2024). AI Jobs Barometer 2024.
- International Labour Organization (2025). Generative AI and Jobs: A Refined Global Index of Occupational Exposure.
- Frey, C. B. & Osborne, M. A. (2013). The Future of Employment: How Susceptible Are Jobs to Computerisation? Oxford Martin School. (Historical anchor only.)
Next ASHE release lands and a figure changes? We will update this piece. Honest salary intelligence means saying when the data has moved.
Topics: ai, future-of-work, uk-salaries, career-risk, ons-ashe-2025