They may have gotten their wish. Of jobs that could be done remotely, 78% of U.S. work locations are currently either remote or hybrid, according to Gallup data, up from 40% in 2019. Meanwhile, fully on-site roles went from 60% of placements in 2019 to 22% this year.

But for every millennial or Gen Xer happily able to take calls in sweatpants Mondays and Fridays, mounting evidence suggests that working in your sweatpants is the real reason, not AI adoption, behind the plunge in entry-level hiring halfway through the decade.

New research from the Federal Reserve Bank of New York puts numbers on the dynamic. Researchers found the unemployment rate for college graduates younger than 29 climbed from 3.1% to 3.7% over the past nine years. Over the same period, unemployment among more experienced college graduates older than 29 actually ticked down, from 1.9% to 1.8%. 

The divergence traces back to “remotable” fields—software engineering, financial analysis, and other white-collar roles. In jobs that require physical presence, like nursing, the age gap spiked briefly in 2020 then normalized. In remote-eligible work, it never did. Remote work could account for as much as 64% of the overall rise in youth unemployment since the pandemic, the researchers found.

Recent graduates aged 22 to 27 are currently dealing with an unemployment rate of 5.6% as of March. It’s higher than the general unemployment rate (4.2%), and well above the share of degree-holders of all ages without a job (3.1%). Many would-be white-collar workers have designated generative AI adoption across U.S. firms as the scapegoat for a lack of entry-level work opportunities. 

The same economists behind the new Fed research recently published another paper for the National Bureau of Economic Research, focusing on productivity among software engineers at a large U.S. firm. The researchers found that while remote work can boost output among experienced workers, it can be to the detriment of younger engineers. 

Feedback on coding work increased 18.3% when workers were in the office, improving the quality of output, according to the paper. Younger workers disproportionately benefited from in-person mentorship and feedback sessions, while periods of flexible work at their company had “scarring effects” on young graduates’ development.

Separate researchers are increasingly aligning on this thesis—that the same privilege workers staged company walk-outs to protect only a few years ago—is at least partially responsible for the wave of youth unemployment and underemployment sweeping the U.S. Firms that operate on distributed teams have become less willing to hire young workers requiring mentorship, economists say, and are happy to keep hiring older and safer staff instead.

Remote work villain

A closer look at who is and who isn’t finding jobs points to remote work as a powerful variable. Comparing unemployment trends across “remotable” occupations, such as software engineering or financial analysis, with occupations that rely on physical presence like nursing, the Fed research found the entire increase in relative youth unemployment boils down to remotable fields. Hiring rates have mostly normalized for physical roles. Nursing, for example, has been one of the labor market’s bright spots lately.

Other research is coming to a similar conclusion. A working paper by economists at the London School of Economics and the University of Oxford, published last month, scanned hundreds of millions of hiring records and job postings in the U.S., U.K., Canada, and Australia between 2017 and 2025. They found that while entry-level hiring has indeed plummeted—between 14% and 29%, depending on which country—senior hiring has risen 5% to 21%.

Remote work appeared to be a key culprit behind the discrepancy. Companies that publicly announced strategy shifts towards working from home or hybrid models early in the pandemic are now more likely to staff senior roles and older workers, with fewer entry-level roles available.

The authors do note that the companies most adapted to offer flexible work options are also more likely to employ for roles that are more easily automated by AI, suggesting some correlation might be at play. 

But both the working paper and the new Fed research point out that the age gap in hiring predates the mass diffusion of AI tools. Over the past few months, companies that have pinned layoffs and muted hiring on automation have been accused of so-called AI-washing, blaming technology for headcount decisions that were likely to happen anyway. Economists have so far been at pains to find evidence that AI adoption is directly responsible for unprecedented shifts in the labor market, finding that its impact so far has been mostly similar to that of the Internet or computers: disruptive, but not apocalyptic.

“There is zero evidence of job losses because of AI,” Torsten Slok, chief economist at investment firm Apollo, wrote in a blog post last week. Citing employment data that has held steady in recent months, Slok said the push for AI adoption might actually be raising demand for jobs as firms hire more engineers and AI experts.

Employers might see AI as a convenient justification for their hiring decisions, though from an economics’ standpoint, the remote work factor might have more to stand on. 

That AI has taken the brunt of blame is likely small reassurance to new graduates struggling to advance in their career. Young workers in general are likely aware that the popularity of remote work is mostly to their disadvantage, as a Gallup poll last year found Gen Z to be the age group least likely to prefer a fully remote workplace setup, citing in part the lack of interaction with coworkers.