Job growth in industries 'exposed to AI' sees rapid drop

The age of AI could tentatively be causing a realignment in the UK labour market. A new study suggests that job postings with higher exposure to AI have grown at a much slower pace since 2022.

Since the hype around artificial intelligence first leaked into public discourse in late 2022, there has been an insistence that generative AI was not going to render all humans unemployed – but rather it would necessitate a more benevolent sounding “transition”. Changes in working practices of some jobs could cause short-term skill mismatches, which would lead to professionals from impacted professions being funnelled into other lines of work. Going into detail of which roles may be for the axe first, however, the benevolence soon fades.

This process has often been cited as primarily being set to impact creative roles like authors, translators, graphic designers, along with IT support technicians and legal professionals, due to the text-based and visual nature of their work. Those lucky workers whose roles would be “significantly impacted” by generative AI would be able to seek new fulfilment in the retail, customer services, hospitality, construction, and manufacturing sectors – areas to experience ‘minimal impact’ from the technology. In broad terms, this would see an inversion of the famous, 1950s-style vision of the future, where robots take on all the gruelling work to free humans up to make artistic, philosophical or scientific advances. 

Judging by the findings of PwC’s latest Global AI Jobs Barometer, this could already be happening – though the firm seems decidedly upbeat about that prospect. Claiming that “AI is sprinkling stardust on UK employees and sectors that are best able to use the technology”,  the study suggests that – as UK unemployment hits its worst levels in four years – job vacancies are growing fastest in sectors where generative AI has the least to offer.

According to PwC, there are the same number of job postings for the 25% of industries most exposed to AI, as there were in 2012 – while these vacancies have declined steeply since 2022. Meanwhile, the same time period has seen the 25% least exposed roles boom to three times their 2012 number.

But what are these jobs? In keeping with earlier forecasts, the least exposed jobs include things like food preparation assistants (over 120% growth since 2019) cleaners and helpers (79% growth) and personal care workers (100% growth). Meanwhile, even as teaching professionals buck the trend (60% growth since 2019), other professions such as information and communications technology work have seemingly entered terminal decline (more than a 20% fall since 2019).

PwC also claims the advent of AI is causing a fall in education and training requirements across all fronts. The top 50% most exposed roles have seen their degree requirements fall since 2019. The study suggests while 64% of these roles required a degree five years ago, in 2024, that had fallen to 56%. Such a decline might suggest that employers are using AI as a means to undercut skilled labour – reducing its bargaining power and pay, by suggesting roles could simply be filled by an AI, and one or two quality-control staff.

Umang Paw, chief technology officer at PwC in the UK, commented, “There are still many unknowns about AI's potential. By highlighting a correlation between industries that are using AI the most and revenue growth, job vacancies, and wages, our findings offer a signal of what AI could deliver… In the Intelligence Age, the fusion of AI with technologies like real-time data analytics—and businesses broadening their products and services—will create new industries and fresh job opportunities.” 

Whether the technology is actually delivering on these fronts might be taken with a grain of salt, however. While proponents of the technology are willing to conflate falls in certain lines of work with AI filling those roles, as the time-frame aligns roughly with the AI hype which emerged in 2022, they are less willing to do the same when it comes to the performance of the global economy in those years – now said by the World Bank to be at its worst since the 1960s. While it might benefit some consultancies to suggest their predictions and partnerships for AI are finally bearing fruit, then, it might pay more not to draw attention to that apparent cause and effect.