From translators to masseuses: Meet the workers most and least at risk of being replaced by AI

Entry-level workers are facing a “job-pocalypse”, a new study has warned, as companies prioritise artificial intelligence over new hires.

The British Standards Institution polled more than 850 bosses in Australia, China, France, Germany, Japan, the UK, and the US – and the results make for sobering reading.

Some 41% of business leaders said that AI is allowing them to make redundancies, while nearly half (43%) expect to reduce junior roles over the next year due to efficiencies in artificial intelligence (AI).

“The picture that emerges is a risk of ‘Generation Jaded (jobs automated, dreams eroded)’” warn the report authors. “A cohort already impacted by disruption from Covid-19 to their schooling and adolescent development, now facing an uncertain employment future with potentially little opportunity to hone their skills or build industry knowledge.”

But there is good news – at least for embalmers, floor sanders and roustabouts. According to another new study these are the occupations least likely to be impacted by AI.

Analysing more than 200,000 anonymised conversations between users and Microsoft Bing Copilot, researchers from Microsoft have determined an “AI applicability” score for dozens of different jobs.

It’s bad news for interpreters and translators – according to the study, AI can assist with 98% of the occupation’s core tasks. Historians follow at 91%, then writers and authors (85%), sales representatives (84%), journalists (81%) and passenger attendants (80%).

“Knowledge aggregation and writing – anything that involves the treatment of knowledge in an administrative sense appears to be relatively vulnerable,” says Andrew Rogoyski of the Surrey Institute for People-Centred Artificial Intelligence.

“That doesn’t straightforwardly mean that they are most likely to be fully displaced, because humans adapt, but AI will shape the ways these workers work and the availability of employment.”

Read more:

• Big tech wants to mine human creativity for AI training. We say no
• Is AI pushing jobless young graduates to the far right?
• How do we stop AI becoming an ‘engine of inequality’, and make it a force for good?

By contrast, jobs requiring physical labour, machine operation, or context-specific manual skills rank lowest in AI applicability. This includes dredge operators, bridge and lock tenders, water treatment plant operators, logging equipment operators, nursing assistants and massage therapists. For many of these “lowest risk” jobs, the reported AI applicability is 0.00 (i.e. rounding to 0%) or extremely close to zero. Embalmers and dishwashers are also, it seems, pretty safe.

So, should every translator in the country retrain in floor sanding?

Not necessarily, says Rogoyski. The study is “very static in their treatment of human skills”, but humans are actually very flexible.

“The study analyses what is currently meant by, you know, a road grader or a knowledge worker in some form,” he said “So it doesn’t really recognise that humans adapt and change with the advances in technology.”

On the flipside, the ‘safe’ jobs might not be safe forever.

“We may regard that plumbers and embalmers and so on, dredgers are safe jobs,” said Rogoyski. “But you know, advances in robotics are happening at pace, and those physical jobs may be very displaceable in five years’ time.”

Still, the risk of mass redundancies is hard to ignore. Last year the think tank IPPR warned of an AI-driven ‘job apocalypse’ without strong government action. In its worst-case scenario, 7.9 million jobs disappear with no economic gain. In the best case, no jobs are lost and GDP grows by 13%.

We’re at a sliding doors moment, where government policy is intersecting with rapid developments in technology. At present, it’s hard to predict which way it will go.

One trend is clear: the young are most at risk. Normally, late summer marks the shift from dissertations to first jobs, but this year the supply of entry-level roles has shrunk. According to Indeed, graduate postings fell by a third over the past year. AI is partly to blame, says Rogoyski.

“Companies looking to make immediate cost savings by not hiring or getting rid of firing low skill, low experience workers – that typically targets the young. Graduate employment is slowing down, for example, not just in the UK, but around the world. And a few recent studies are suggesting that actually this might be due to early effects of generative AI,” he adds.

“But how are people to learn their craft and their trade? You jeopardise the talent pipeline.”

The numbers bear him out. Entry-level vacancies have collapsed since generative AI went mainstream: Job search site Adzuna recently reported a 32% fall in junior and graduate roles since late 2022, with graduate postings at their lowest level since 2020 .

In tech, the drop is even steeper – research by Ravio finds entry-level hiring across Europe down by more than 70% in the past year. And a survey by Prospects & Jisc career guidance consultancy suggests the anxiety is filtering through to graduates themselves, with one in ten already changing career plans because they fear AI will make their chosen path obsolete.

By culling a graduate recruitment pathway, it may well do so.

“You can’t just delegate everything entry level to an AI. If you displace, let’s say, a sort of entry level bookkeeper with an AI. That bookkeeper may have gone on to become a sort of high flying accountant, or a financial director of a FTSE 100 company or something, but that track is no longer there.”

There’s another risk looming over the job market, anyway.

There is a growing threat that the Artificial Intelligence bubble could burst, the Bank of England has warned – an outcome that could trigger a recession.

Frenzied hype about AI’s potential has led to massive increase in valuations, with companies such as OpenAI now worth $500bn (£372bn). But is it sustainable?

The Bank of England has warned there is a growing risk of a “sharp market correction,” with valuations plunging. It could have serious implications for Rachel Reeves’ growth forecasts

“The risk of spillovers to the UK financial system from such global shocks is material.”

But before that happens, Big Issue spoke to the profession most and least at risk from AI – according to the Microsoft study – about how their jobs are already being impacted.

Translator: the most impacted by AI

Beth – not her real name – is an Ireland-based translator in Spanish and Portuguese. She’s been working in her field for 20 years.

How has AI impacted your work?

Yes, it has. Everything was going great before. I lived between London and home to look after my mum, and translation really suited me – not just the work itself, but the flexibility. But suddenly things have whittled down. My work is probably about 50% of what it was two years ago. Nobody’s written to say “we’re using AI”, but agencies that once gave regular work have become platforms, or been bought by American conglomerates. Now most jobs are bottom-line, first-come-first-served. I got a job offer the other night, couldn’t even take it because the computers were down, and by the time I logged on, it had been snapped up. Super competitive.

Are colleagues using AI?

I think it’s a mixed bag. People think it’s overhyped. Yes, AI can do certain things, but language isn’t just an equation. You can’t break it down like that – it’s about nuance, culture, meaning. Some people use it for research, maybe summarising, but it’s still very reductive.

What can AI not do?

Well, I never use it. It misses the subtleties – empathy, context, cultural understanding. If you throw hospital reports or legal documents at it, there can be a mismatch. Sometimes AI makes things up. With translation that’s a serious risk.

Do you think human translators will always be needed?

I’d like to think so. But the economy isn’t brilliant. Businesses want to cut corners. Sometimes it’s cheaper to machine-translate. It’s about what people value. If they want quality, human translators are still essential.

Massage therapist: the least impacted

Danny Morgan is an osteopath and massage therapist. He runs Body Function Clinic, overseeing 11 therapists taking about 400 appointments a month.

Is AI a concern for your profession?

There’s a lot of talk about AI, but we don’t feel too scared. The human element – hands-on work, really knowing a patient – can’t fully translate to AI. Some physiotherapy platforms try video consultations, telling patients to squat or do exercises, then giving generic routines. It’s like turning them into mannequins. That hands-on skill takes years to develop; AI isn’t close to that.

Do you use AI in your clinic?

Yes, operationally. We use software to record sessions and auto-generate treatment notes. It saves us hours on admin. But I’ve deliberately avoided AI answering phones – people, especially older patients, want human interaction. Healthcare is built on relationships and trust. AI can assist, but it can’t replace that connection.

Do you think AI could ever replace massage therapists?

I don’t think so. It can help with tasks like writing letters or managing workflow, but people want empathy and real connection. Recovery often depends on the practitioner–patient relationship, and AI won’t get that.

Would you recommend a young person train in therapy because of AI?

Yes, but it’s hard work and requires good communication skills. Don’t pick it just to avoid AI. The shelf life of your average physio is about seven years, so it’s not for everyone.

Do you have a story to tell or opinions to share about this? Get in touch and tell us more. 

Reader-funded since 1991 – Big Issue brings you trustworthy journalism that drives real change.

Every day, our journalists dig deeper, speaking up for those society overlooks.

Could you help us keep doing this vital work? Support our journalism from £5 a month.