IMPARTIAL NEWS + INTELLIGENT DEBATE
Account
Already this year the number of posted graduate entry-level jobs is down by around 10 per cent. Is AI to blame?
In January Keir Starmer staked the UK’s future on AI. In a keynote speech, the Prime Minister promised significant government investment, friendly regulations, and more to attract AI companies to the UK – in a big to boost the UK’s struggling economy. Starmer said he believed AI could deliver investment, growth and high-skilled jobs to the UK.
But even as he promises the new technology will create jobs, people in multiple professions are warning that AI is doing the opposite – threatening entry-level jobs in careers including law, consultancy and communications.
Despite repeated efforts by the Government to encourage schools to teach coding – and young people to get into IT as a career of the future – it’s jobs in software engineering that are being hit first by the rise of AI, as ChatGPT and its rivals are already capable of writing working code in seconds.
Tech leaders are saying this openly: the venture capitalist Marc Benioff, who is CEO of the software company Salesforce, announced he would not be hiring any software engineers throughout 2025, “because we have increased the productivity this year with Agentforce and with other AI technology that we’re using for engineering teams by more than 30 per cent”.
Similarly, Meta CEO Mark Zuckerberg announced that within the year that Facebook was “going to have an AI that can effectively be a sort of midlevel engineer that you have at your company that can write code” – before announcing that 5 per cent of all Meta employees would be losing their jobs.
The result is that UK graduates looking for roles in IT and software are finding it almost impossible to find roles, despite believing they had chosen a degree that would significantly boost their employment prospects.
“I completely lucked in to my first role,” said one IT worker, who like other working professionals in this piece is not being named as he was not authorised to speak about internal matters to the media.
He says “companies need things done yesterday, though, so don’t want to take the time” to train people, now that AI is an easy alternative that can automate the straightforward tasks entry-level IT workers would otherwise do.
But software engineering might just be the canary in the coalmine: graduate jobs are getting harder to get across the economy – and while AI might only be one of many factors, it does suggest the technological progress on which the Government is depending could come at a cost for those just entering the workforce.
Figures collated by the Institute of Student Employers show that competition for graduate jobs in 2024 was much fiercer than even 12 months before, with a typical graduate trainee role having 140 applicants, up from 88 in 2023. In the early months of this year, the number of posted graduate entry-level jobs is down by around 10 per cent across the country and 20 per cent in London.
“When I joined my firm on the grad scheme, each partner had their own executive assistant and directors had access to them too,” said a millennial employee of one of the Big Four accountancy firms.
Earlier automation, such as transcription software, research tools that scan the internet and more, meant that many of those roles disappeared. “Now only partners have access to assistants, and they have to share one between two or three partners,” he said, before explaining that the disappearance of office roles was now expanding to include analysts, a role typically filled by graduate trainees.
“If Microsoft Teams can summarise the meeting – including capturing actions etc – do you really need to have an analyst in the meeting at all?” he said. “If generative AI learns to make good-looking PowerPoint slides then analysts really are in trouble…”
Several of the UK’s leading white-collar industries, including consultancy and law, still bill their clients by the hour, charging a fee related to the amount of time different staff or partners spent on a project that differs according to their seniority.
Having a lot of junior staff like analysts has traditionally been an asset to the Big Four accountancy businesses, because they could stack up lots of billable hours and thus be a profit centre for the firm – especially because clients were less likely to challenge time spent by junior staff on a project versus their senior colleagues.
“Now, clients are becoming less willing to pay for the analysts,” the consultant continued, noting that “already we’re seeing there is less demand” for the more basic tasks that his junior colleagues would carry out – because clients now expect that these basic research and collation tasks can be done much more quickly and cheaply by AI, and then reviewed by someone more senior.
“That means we might end up going from, ‘we want lots of analysts on the project to boost our margin’ to ‘we will accept a number of analysts so they get experience so they can do the more complicated things in the future’,” he concludes. “They’re becoming more of a ‘cost of doing business’ rather than a big profit engine.”
This concern is mirrored across other professions, where people with more experience – millennials in their 30s and 40s – are worried that even where Gen Z graduates can get into their industry, the way AI has automated some of the tasks junior staff would do means they don’t have chance to pick up vital skills.
“The big problem with AI for communications and public affairs is that it’s automating so many processes that younger people in teams previously did, which taught them how to be good at their jobs,” one senior political communications worker – who has worked for the government and in private consultancy – explained.
“It’s not just that the jobs are going to go – there will still be some jobs – it’s that the kids coming into those jobs, they aren’t learning as much as they think they are, because it’s possible to automate so much of it.
“I see this in in politics, with people saying, ‘Oh, you know, we can get an AI to analyse this’, and it’ll go through, like, 450 parliamentary questions in a minute and a half and have an answer,” she continued. “It’s like, okay, but if you as the intern had to do that 20 years ago, you learned so much by doing it, and your ability to take that knowledge and progress with it was better.”
The result, she warned, could eventually prove disastrous for British politics. “It makes me worry about the kind of advisers we’re going to produce in the long run,” she said.
Recent research by Adzuna utilised data from Goldman Sachs and the Organisation for Economic Co-operation and Development (OECD) to identify those jobs safer from AI, then ranked them by offered salary. The top five were:
Dermatologists (average salary £86,229)
Surgeons (£80,000)
Psychologists (£76,400)
Paediatricians (£74,600)
Dentists (£70,000)
Other jobs unlikely to be replaced by AI are priests, hairdressers, roofers, plumbers, carpenters, builders, labourers, lifeguards, yoga teachers – and crane drivers.
Also AI-proof, for now at least, is any job that requires human reasoning and judgement for example: nurses, therapists, social workers, midwifes and museum curators, compliance officers, therapists and investigative journalists as well as marriage counsellors, physiotherapists, vets – and pest control technicians.
The UK’s prestigious professions are organised in “pyramids”: at the top are partners, who own a share of the firm and thus get a proportion of its profits every year. At each tier below there are progressively more junior staff, in increasing numbers. Ultimately, for every partner there are several people hoping one day to net one of the top spots.
Analysis by Thomson Reuters shows that in the UK’s legal sector this traditional structure is changing. In 2005 to 2009, the bottom rung of the legal profession, associates, made up around 45 per cent of employees. By 2020-2024 this proportion had dropped to 40 per cent, suggesting that the influx of new blood into the sector is diminishing.
Richard Tromans, founder of Artificial Lawyer, the legal AI information site, says the trend so far is mostly explained by law firms holding onto more senior associates for longer – but he does think AI will change the profession. “As legal AI is steadily brought into billable work, then law firms may find they need less junior lawyers.”
“This could prove complicated for a business model which up until now has been based on leveraging a lot of junior talent,” Tromans explains. “The biggest issue around AI and junior lawyers in the future is: how will they learn to be good lawyers if a lot of entry-level work has been automated?”
Automation has been a threat to jobs ever since it existed: the Luddites famously smashed up mechanical looms in the 19th century. Robots have replaced most skilled manufacturing jobs, while word processing and email replaced many secretarial and PA jobs.
Now AI has arrived, it may be the turn of white-collar workers to be replaced, with the lucky few graduates who get in having a harder time learning on the job, too – but Tromans thinks AI might be able to help with that, if it’s used right.
“One idea is to create AI-driven legal simulators that can help junior lawyers to learn outside of any billable work,” he suggests. Perhaps it’s not all bad news for recent graduates, after all?
Impartial news + intelligent debate
All rights reserved. © 2026 Associated Newspapers Limited.
Leave a Reply