Research from the London School of Economics, Oxford University and the New York Federal Reserve found that when economists control for the rise of remote work, the apparent effect of AI on entry-level jobs nearly disappears.
The study looked at hiring patterns across sectors that have adopted AI tools most heavily, comparing them against sectors where remote working arrangements expanded fastest over the same period.
Where earlier analyses had assumed a direct link between AI rollout and a decline in junior hiring, the researchers found that adjusting for remote work explained away most of that apparent relationship.
Remote work factor
The mechanism appears to be structural rather than technological: as more entry-level work has shifted to remote or hybrid arrangements, the traditional pipeline of junior roles, often built around in-person mentoring, on-the-job training and physical proximity to senior staff, has been disrupted independently of any AI deployment.
Fewer junior employees are being hired into roles that once relied on sitting alongside more experienced colleagues, absorbing skills informally rather than through structured training programmes.
That erosion, the research suggests, would likely have occurred even in the absence of generative AI tools, simply as a consequence of remote and hybrid work becoming permanent fixtures of corporate life.
Challenging the narrative
This finding cuts against a prevailing assumption among commentators, executives and policymakers that AI is the primary force reshaping the bottom rung of the career ladder.
That assumption has taken hold partly because the two trends, AI adoption and a documented decline in entry-level hiring, have unfolded over roughly the same timeframe, making the two easy to conflate.
Correlation between AI adoption and entry-level job losses does not necessarily mean causation, and disentangling the two requires the kind of statistical controls applied in this research rather than casual observation of parallel trends.
If remote work is in fact the bigger driver, then policy responses and corporate hand-wringing aimed squarely at AI may be targeting the wrong problem, leaving the actual cause of the hiring squeeze unaddressed.
Wider context
The suggestion that AI will not simply take people's jobs, but that those who fail to learn how to use it may lose ground to those who do, remains a widely held view among technology commentators and business leaders.
These findings add an important layer of nuance to that position, suggesting some of the anxiety currently directed at AI and its effect on hiring may be misdirected at the wrong culprit entirely.
For employers and graduates alike, the implication is that fixing the entry-level jobs problem may require rethinking how remote and hybrid work is structured, rather than simply pausing or reversing AI adoption.
