TL;DR

Office for National Statistics data indicates AI adoption is not costing jobs at any scale in British firms. Adoption among firms with 10 or more employees rose from around 12% in late 2023 to around 35%, yet just under 7% of medium-sized businesses reported cutting headcount as a result. Around half of businesses reported no impact on headcount at all. Training was the common response: among businesses citing a lack of AI expertise as a barrier, around 62% said they were building skills by training or retraining existing staff. The ONS figures were flagged in analysis by hosting provider Fasthosts, reported by CFOtech UK.

Adoption skews to scale. Almost half of businesses with 250 or more employees use at least one AI technology, against 28% of those with up to nine staff. Improving business processes was the most common internal use, particularly above 50 employees — administrative work, workflow management and routine customer interactions rather than whole-role replacement.

The UK is not an outlier in the data. Research cited by the St Louis Fed found 36.3% of UK workers use generative AI at work, ahead of Sweden, the Netherlands and Germany, against 43% in the US.

Separate US research points the same way from a different direction. A note by former Bureau of Labor Statistics commissioner Erika McEntarfer and colleagues at Stanford’s Institute of Economic Policy Research found employment trends in occupations highly exposed to AI are “fairly stable”, and that online job postings for software developers have grown faster than for other occupations over the past year. The authors offer two explanations: the share of tasks AI can profitably speed up may be small relative to all economic tasks, and adoption remains shallow — the US Census Bureau estimates around 20% of companies use AI, and a McKinsey survey found most are still piloting.

There is one clear exception. Employment of 22- to 25-year-olds in software development and customer service has been falling relative to older cohorts since ChatGPT launched in November 2022, though isolating the cause is difficult.

Looking Forward

Two independent evidence bases now say the same thing, which makes the displacement narrative harder to sustain on current data — and also harder to dismiss, because neither dataset can see far ahead. Hiring is not expanding either: just 1% of the smallest UK firms and 1.2% of the largest said AI directly increased headcount, so this is job content changing rather than jobs appearing or vanishing. The entry-level finding deserves separate attention, given the Burnham government’s youth training overhaul. Slower adoption buys policymakers preparation time; it is considerably less welcome news for anyone holding AI valuations.