It’s Remote Work, Not AI, That Is Stalling Graduate Hiring, Study Shows

A rigorous analysis of more than 400 million job advertisements and 240 million hires across the United States, Britain, Canada and Australia has overturned the prevailing assumption that generative AI is the primary force hollowing out entry-level recruitment. The study, by researchers at the London School of Economics and the Allison Institute of Technology, finds that the shift to remote work—and the attendant difficulties of mentoring, monitoring and integrating junior staff from a distance—is the dominant factor suppressing graduate and early-career openings. Controlling for AI exposure, occupations with high remote-work feasibility saw entry-level hiring fall by four to five percentage points more than those that cannot be done from home. Fresh data from the New York Federal Reserve underscore the distress: the unemployment rate for recent university graduates has climbed to 5.7 per cent, well above the overall 4.2 per cent, and entry-level posts in the US are down 29 per cent compared with pre-pandemic levels.

The findings land with unusual force because they arrive just as Sam Altman, the chief executive of OpenAI, has publicly walked back his earlier fears. Addressing a Commonwealth Bank of Australia conference in Sydney, Altman confessed he was “pretty wrong” about the social and economic effects of the technology his company unleashed. “I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened,” he said. That candid admission, from the corporate figure most identified with the AI boom, deepens the intellectual puzzle: if not AI, what is closing the door to the young?

The answer provided by the cross-country study—remote work—is, however, only part of the story. Viewed from Jakarta, where local media have seized upon a Fortune magazine feature declaring the diploma’s declining value, companies are rapidly abandoning the old credentialist model and demanding practical, higher-order skills from even their youngest hires. The Fortune analysis notes that entry-level positions now often demand what were once considered mid-career competencies. Meanwhile, the LSE-Allison researchers acknowledge that AI is quietly rewriting job content: routine tasks once assigned to new graduates are being automated, and firms are expecting juniors to shoulder more complex responsibilities earlier. Thus, the entry-level role is not merely vanishing; it is being redefined in ways that make it harder to access for those without demonstrated, often non-academic, expertise.

This pattern is vividly illustrated in European banking. Sweden’s financial supervisor reports that 84 per cent of domestic financial firms already use generative AI, yet only 22 per cent have AI systems in production within their own IT environments. Across the European Union, according to the European Banking Authority’s 2025 survey, 92 per cent of banks are now AI users. The transformation underway, analysts in Stockholm note, is less a story of headcount reduction than a deep reorganisation of work: banks are not simply cutting junior positions; they are unbundling tasks, automating repetitive processes and demanding a hybrid skill set from the employees who remain. The challenge for managers is not whether to deploy AI, but how to make it effective inside labyrinthine legacy systems built over decades.

The immediate implication for policymakers and educators is that a narrow fixation on AI as job terminator misses the more prosaic but powerful impact of working from home on young professionals’ career formation. If remote work impedes the transmission of tacit knowledge and slows network-building, then companies that continue to hire but fail to provide in-person mentorship may be incubating a future productivity deficit. At the same time, the creeping automation of routine tasks suggests that even if a “jobs apocalypse” has not arrived, the pathway from education to fulfilling employment is narrowing and steepening. Altman’s relief might be short-lived if the combined effects of dispersed work and task automation leave a generation of graduates caught between the disappearance of old entry points and the rising threshold of the new ones.