AI isn’t destroying entry-level jobs. It’s changing them

{"text":[[{"start":7,"text":"In 2023, researchers studying AI found that while large language models were surprisingly good at simple tasks carried out by employees at Boston Consulting Group, they struggled when given more complex work."}],[{"start":19.1,"text":"In the years that followed, as generative AI sliced through the basic repetitive work traditionally assigned to junior employees in professional services, many companies decided to pause entry-level recruitment. "}],[{"start":32.900000000000006,"text":"This year, the prospects for entry-level workers appear brighter. Some companies have realised that young staff can replenish the company hierarchy and bring native AI fluency to the job. "}],[{"start":43.550000000000004,"text":"An April survey by the National Association of Colleges and Employers showed US companies expected to hire 5.6 per cent more new college graduates this year, compared with last year’s forecasts of a flat job market for the class of 2026."}],[{"start":58.650000000000006,"text":"Many young job hunters are still desperately fishing in a shallow pool of entry-level opportunities. But companies in sectors ranging from consulting and law to engineering and fashion are already updating their hiring practices, training programmes and working culture in readiness for the coming AI era."}],[{"start":null,"text":"

"}],[{"start":76.5,"text":"US employers making the biggest AI investments, including software and media companies, increased employment by about 10 per cent after adopting AI tools, according to a study by jobs data company Revelio Labs and fintech Ramp. "}],[{"start":91.75,"text":"The heaviest adopters are also hiring more entry-level employees. For these young workers, human skills including critical thinking and professional judgment are catching up with technical abilities in coding in the list of required attributes. "}],[{"start":105.75,"text":"The adoption of AI in companies is also changing the nature of the work that new recruits do. “We are starting to see encouraging signals that entry-level work is moving from grunt work to real work,” says Aneesh Raman, chief economic opportunity officer at careers site LinkedIn. "}],[{"start":123.45,"text":"But as companies move towards an AI-augmented future, some experts worry that the removal of mundane repetitive work could starve early-career staff of the experience that they might one day need when overseeing hybrid teams of humans and AI agents. "}],[{"start":139.85,"text":"If companies can master AI at the lower levels, the benefits will rise up through the hierarchy, experts say. “Entry-level work is representative of where all work is going,” says Raman of LinkedIn. “If we solve [the challenges] for this category of work we are solving for all work.” "}],[{"start":157.35,"text":"Redesign jobs"}],[{"start":159.85,"text":"Candidates for starter roles in the most AI-exposed industries are now expected to show a mastery of the skills traditionally demanded of more seasoned staff, such as data-driven decision-making and people management."}],[{"start":null,"text":""}],[{"start":172.75,"text":"A recent survey by PwC of more than 1bn job adverts on six continents describes this “seniorisation” of entry-level jobs, finding that since 2019 there has been a 35 per cent increase in such postings. "}],[{"start":187.55,"text":"Salesforce, a cloud-based software company, is taking on 1,000 recent graduates and interns this year, more than it hired in 2025. They will be working on AI projects across the business, says Paula Goldman, the company’s chief ethical and humane use officer. “These roles are being imagined differently; they’re being imagined as we are going through a particular transformation and we think that making a big bet on this kind of talent is going to accelerate [the change],” she adds."}],[{"start":213.85000000000002,"text":"Nobody knows for sure what the introduction of more AI-enabled work might do to company hierarchy. Ravi Kumar, chief executive of consultancy Cognizant, believes middle managers, traditionally the measurers, orchestrators and co-ordinators of work at large companies, will have to become “player-coaches”, helping to guide juniors. "}],[{"start":null,"text":""}],[{"start":234.70000000000002,"text":"Eleanor Lightbody, CEO of Luminance, which develops AI for the legal profession, says these middle layers could be squeezed out altogether, “because we are going to hire more juniors, who are really going to understand how AI works, and more seniors [are] staying in the business because [by using AI], they can be more productive and have more capacity”. "}],[{"start":255.05,"text":"The challenge for job seekers is that it is hard to find entry-level jobs that allow them to develop the higher proficiency that seniorised roles now require."}],[{"start":265.8,"text":"Hire for ‘power skills’"}],[{"start":267.6,"text":"When they imagine a new car, young designers at Italian studio Pininfarina still start by picking up a pencil. "}],[{"start":274.6,"text":"“The important thing is that [they] pass through the real original creative process, which is the hand on the sketch,” says Marco Busi, senior vice-president for mobility at the Turin-based group, which has designed vehicles for manufacturers from Peugeot to Ferrari. "}],[{"start":289.55,"text":"Just as computer-aided design put paid to the drawing board in the 1980s and 1990s, AI is becoming part of the latter stages of Pininfarina’s design process. Generative AI already fine-tunes company presentations and one day it should help Pininfarina produce 3D model vehicle prototypes. "}],[{"start":308.65000000000003,"text":"But forcing the newest recruits to sketch in the first instance helps to avoid “bringing bias into the creativity”, Busi says."}],[{"start":317.05,"text":"Reliance on AI is increasing the demand for distinctively human “soft skills” — or “power skills”, as some are now calling them — such as creativity, empathy, judgment and networking ability. "}],[{"start":328.8,"text":"PwC’s jobs report found that new tasks added to job adverts for AI-exposed roles were two and a half times as likely to call for such capabilities. "}],[{"start":339.05,"text":"The problem with large language models, says Gary Marcus, professor emeritus at New York University, “is that we have young people who get dependent on these things, they learn nothing while they’re using them and they don’t develop critical thinking skills”. This means they are worse prepared to oversee the next generation of trainees and apprentices and are less able to detect flaws in the output of AI itself. "}],[{"start":null,"text":""}],[{"start":362.05,"text":"The need for critical thinking is strongest at companies where the consequences of AI-induced error are highest."}],[{"start":368.15000000000003,"text":"Daniel Lee, chief exterior designer at Pininfarina’s Shanghai studio, says AI can lead young designers to “jump to the conclusion”. “Back in the day, you would make mistakes on the proportions or the line work. You would need to practise constantly to get it perfect . . . But now designers wouldn’t see that because it’s being hidden by the colour or the background [of an AI-rendered image].” "}],[{"start":391.55,"text":"That can lead to basic errors such as omitting a radiator vent on the visualisation of an internal combustion engine-powered car."}],[{"start":399.25,"text":"The snag is that soft skills are hard to assess and develop. Educational institutions may be overestimating how good a job they are doing. Some 78 per cent of higher education leaders think their graduates meet employer expectations, according to a recent survey by education company Pearson and Amazon Web Services. But more than half of employers polled said they still struggled to find AI-ready graduates. "}],[{"start":425.7,"text":"“We can’t have parallel paths where institutions are on one path and enterprises are on another. We have to come together,” says Pearson chief human resources officer Ali Bebo."}],[{"start":435.75,"text":"Expand who and where you recruit"}],[{"start":438.6,"text":"Economist Ben Zweig doubts the recent decline in entry-level jobs is directly caused by the advance of AI, or that it is likely to persist. Instead, Zweig, who is also chief executive of Revelio Labs, attributes it to the volatile outlook for businesses assailed by tariffs, war, and the threat of inflation."}],[{"start":460.45000000000005,"text":"“It may make sense to take a bet on a younger worker for the longer term but companies make safer bets on older and more experienced workers in times of uncertainty,” he says."}],[{"start":469.70000000000005,"text":"AI has broadened the pool of potential candidates that employers are prepared to consider. The existence of coding agents has renewed demand for humanities graduates and is encouraging employers to source candidates from outside elite universities. "}],[{"start":null,"text":""}],[{"start":485.75000000000006,"text":"Liberal arts students bring to the job their social skills and expertise in interdisciplinary work. Bebo says this “opens up the aperture” for candidates. They no longer “have to be growing up in the coding world, because everybody can code”."}],[{"start":499.55000000000007,"text":"As architecture became more dependent on computer technology, ZHA, the practice founded by the late Zaha Hadid, considered programming prowess a useful addition to candidates’ technical qualifications. Nils Fischer, who runs ZHA’s technology teams and oversees research and development, says the advent of AI means “concepts have become more important than technique”. "}],[{"start":523.35,"text":"In practice, though, recent graduates still face many obstacles in the jobs market. Harvard Business School’s Joseph Fuller, co-lead of the university’s Managing the Future of Work initiative, says one problem is that AI is advancing faster than they can. “The half-life of these [large language] models is distinctly shorter than the time it takes the average worker to get comfortable with the new model,” he says. "}],[{"start":547.1,"text":"That means recent graduates will need actively to update their knowledge of the advancing technology while “companies are going to be increasingly obliged to get good at something they’re terrible at”, adds Fuller — structuring rapid bursts of training for their employees."}],[{"start":562.2,"text":"Use AI as a trainer"}],[{"start":564.2,"text":"AI is not only good at the repetitive tasks on which young recruits typically train. It can simulate the more complicated problems that a new lawyer or engineer might face, allowing trainees to jump to a more advanced level of competence. It can also measure the “AI fluency” of workers, offering them on-the-job training tips. "}],[{"start":583.3000000000001,"text":"In Germany, DHL’s post and parcels business is training AI tools on its official manuals and asking experienced workers to add to or amend the knowledge before they retire. The idea is to build institutional memory that everyone can draw from."}],[{"start":null,"text":""}],[{"start":599.35,"text":"Cognizant, the consultancy, is putting in place an AI “harness” that helps guide new entrants and substitutes for the sort of knowledge new staff might previously have gleaned by “walking the corridors”. Kumar, the company’s chief executive, likens this to a self-driving Tesla programmed with the community knowledge of millions of drivers, not just the individual behind the wheel."}],[{"start":620.2,"text":"Lightbody, CEO of Luminance, believes its AI platform, which is trained on reams of legal documentation, can help young lawyers. “Because people have more access to knowledge faster, they’re capable of doing much more early in their careers and learning much more,” she says. “Rather than waiting to have access to the general counsel and [asking] why they did that in a certain way, all that’s captured in our systems.”"}],[{"start":643.5,"text":"Collaborate — in person"}],[{"start":646.2,"text":"Successful teams pair the deep proficiency of more experienced professionals with the AI and data knowhow disproportionately found among younger staff members, managers say."}],[{"start":657.1500000000001,"text":"“You can have the best proposal using AI [but] if you can’t get it across in a conversation, it’s a problem,” says ZHA’s Fischer, adding that this ability develops through in-person teamwork. In fact, he thinks remote working may have had a more damaging impact than AI on the soft skills of junior colleagues."}],[{"start":676.45,"text":"Only 10 per cent of new employee development at engineering consultancy Arup is based on formal training, says Fiona Cousins, who oversees innovation and technology at the company. Some 70 per cent is learnt on the job and the rest is based on what juniors pick up from other people."}],[{"start":693.2,"text":"Depending on which of the firm’s disciplines they apply for, applicants at Arup still need a thorough technical grounding and a sound knowledge of raw physics. “I’d rather take someone who has been through a proper training in that area and has a reasonable facility in digital and AI” than vice versa, says Cousins. "}],[{"start":712.7,"text":"At Pininfarina, designers now pitch 200 designs with the help of AI, where they used to send just 20. But Lee of Pininfarina’s Shanghai studio says the fundamental question is not “do you use AI?” but “how do you use it and does it fit this brand?”."}],[{"start":729,"text":"There is no technological substitute for core expertise, nor aptitude. “The thing we look for is the one thing that’s very hard to train for,” says ZHA’s Fischer, “which is talent.”"}],[{"start":740.6,"text":"This article has been corrected to clarify that Eleanor Lightbody is CEO of Luminance, not a founder."}],[{"start":754.2,"text":""}]],"url":"https://audio.ftcn.net.cn/album/a_1784293327_7314.mp3"}