The short answer
AI is changing UK STEM jobs most visibly through tasks and workflows, while the national employment effect remains uncertain. Artificial intelligence is changing some work in the UK, including tasks performed in science, technology, engineering and mathematics. The evidence is clearer on shifting workflows and demand for AI-related skills than on the number of jobs that AI has already created or removed. For a UK reader asking whether AI is changing UK STEM jobs, that means describing task-level change with more confidence than net employment effects. No single current estimate tells us the net employment effect across all UK STEM occupations.
To assess AI changing UK STEM jobs, keep three questions separate. This framing also helps when reading AI impact on UK STEM jobs headlines: could AI perform or assist some tasks; are employers actually using it in those tasks; and what happened to hiring, hours or headcount afterwards? Each question needs different evidence.
Exposure, adoption and employment are different measures
AI exposure indices compare capabilities with tasks or abilities associated with an occupation. A high score can identify work that may be affected, but it does not show whether a particular employer has adopted a tool. Adoption evidence describes actual use. Employment evidence records jobs, vacancies, hiring or worker numbers over a defined period.
Even when adoption and employment change occur at the same time, that does not establish that AI caused the change. Demand, interest rates, investment, regulation, outsourcing, organisational restructuring and other technologies can move hiring too. The UK government’s assessment explicitly warns that exposure is not adoption and that observed patterns do not yet establish causality.
Exposure tells us where to investigate. It is not a count of jobs at risk.
What UK projections say about AI-related work
A UK government-commissioned study projects that jobs directly involving AI activities could rise from about 158,000 in 2024 to around 3.9 million by 2035. A broader estimate identifies 9.7 million people in roles where AI may be adjacent to their work. These are modelled projections, not observed 2035 outcomes and not a forecast of 3.9 million new AI-specific vacancies.
The study expects much of the expansion to come from existing jobs taking on AI responsibilities. It distinguishes AI experts, specialists, implementers and general users. This matters for STEM careers: an engineer or research scientist may use AI as part of a domain role without becoming an AI researcher.
The source uses vacancy analysis and modelling commissioned by government; its figures depend on definitions of AI activities and labour-market assumptions. Treat them as an indication of the possible scale of work involving AI, not as a precise count of roles that will exist.
Digital and engineering demand: growth alongside uncertainty
Skills England and the Department for Science, Innovation and Technology identify 30 priority occupations in the digital and technologies sector. Their central scenario projects an increase of about 239,000 workers between 2025 and 2035, with a separate estimated replacement need of about 249,000. The report’s combined demand estimate is approximately 488,000 workers.
These figures cover a defined sector and occupation set; they are not a forecast for all STEM jobs in the UK. The assessment says the projections are illustrative, shows a lower alternative growth scenario, and notes that evidence for AI-related changes in employment levels remains weak. It did not adjust its central projections for AI adoption because evidence was not yet strong enough.
The same assessment describes a move away from some routine coding and testing toward oversight, assurance, judgement and communication. It also identifies electrical engineers, electronics engineers, production and process engineers, and engineering technicians among occupations shared across multiple priority sectors. Those overlaps show why AI and STEM job changes cross sector boundaries.
The government’s Digital and Technologies Jobs Plan separately estimates that about 800,000 roles may need filling from 2025 to 2035 in backed sectors. That total includes projected new roles and replacement hiring; the government explicitly says it is not a target or a forecast of net new jobs. Keep this plan’s scope distinct from the Skills England assessment.
What the evidence says about early-career work
The UK Jobs Plan describes barriers facing people entering digital and technology careers. In cyber security, it reports that the share of job adverts open to candidates with less than one year of experience fell from 25% in 2022 to 16% in 2025. This is evidence about one technology occupation, not all STEM graduate jobs and not proof that AI caused the change.
The UK labour-market assessment also discusses research from the United States that found a decline in employment among younger workers in highly AI-exposed occupations. That finding is international evidence from a different labour market. It should not be presented as a measured UK effect.
A separate 2026 SThree report surveyed 4,126 STEM professionals across the UK, US, Germany, the Netherlands and Japan. It reports views on AI and junior work across its participating markets. Because it is a commissioned industry survey with a five-country sample, its overall findings are not automatically representative of UK workers; country-specific results should be checked in the report before making UK claims.
How to read headlines and claims
- Check whether a number refers to people, jobs, vacancies, tasks, skills or organisations. For any claim about AI changing UK STEM jobs, record the measure and denominator before comparing it with another report.
- Check the geography, occupations, date range and sample. “UK STEM” and “digital technology sector” are not interchangeable populations.
- Separate a forecast from observed data. Projections depend on assumptions and can change.
- Ask whether AI was measured as exposure, workplace adoption or a reported reason for a decision.
- Look for alternative explanations before attributing hiring or job losses to AI.
For an individual role, examine task changes and employer practice rather than relying on a broad occupational exposure label. A useful evidence note names the occupation, geography, period, source method and whether the result is observed, projected or self-reported. That simple checklist prevents a technology capability from being mistaken for a confirmed employment outcome. It is a practical way to assess UK evidence on AI and jobs without treating a projection, survey or exposure score as interchangeable. See our guides to early-career STEM routes and AI skills for STEM work.
Sources and method
When assessing AI changing UK STEM jobs, this page is an editorial synthesis of the linked sources, not an official government view. We prioritise original UK government publications for national projections and labour-market assessment, and keep private surveys labelled as third-party research. Figures are stated with their source’s population and period. Accessed 25 October 2026.
- AI Skills for Life and Work: summary report ↗UK government-commissioned research · Published 2026; research commissioned in 2023 · Accessed 25 October 2026
- Assessment of AI capabilities and the impact on the UK labour market ↗DSIT and AI Security Institute assessment · Published 28 January 2026 · Accessed 25 October 2026
- Sector Skills Needs Assessment: Digital and technologies ↗Skills England / DSIT assessment · Updated 6 October 2026 · Accessed 25 October 2026
- Digital and Technologies Jobs Plan ↗UK Government plan · Published 21 October 2026 · Accessed 25 October 2026
- The STEM Workforce Report 2026: Built to Adapt ↗Five-country YouGov survey, commissioned by SThree · 2026; fieldwork 10 July–1 September 2026 · Accessed 25 October 2026
Frequently asked questions about AI changing UK STEM jobs
Is AI already replacing UK STEM jobs?
Some tasks are being automated or assisted, but current evidence does not establish a net AI-caused loss across UK STEM jobs. Exposure, adoption and employment outcomes must be measured separately.
Does the 3.9 million estimate mean 3.9 million new AI jobs?
No. It is a projection of jobs directly involving AI activities by 2035, and the study expects many existing roles to gain AI responsibilities.
Are software jobs the whole story?
No. Government analysis identifies AI-related work across IT, research, business and finance, while digital and technology occupations overlap with engineering, life sciences, advanced manufacturing and other sectors.
How often is this page updated?
We review the linked evidence when major UK reports or relevant data releases appear. The date near the top shows the last research check; each source also carries its publication or update date.