Remote Work, Not AI, Is Sidelining Recent College Graduates

For years, the prevailing narrative has been that artificial intelligence is coming for entry-level jobs. But new research from the Federal Reserve Bank of New York tells a different story — one that points not to chatbots and algorithms, but to the lasting structural shift in how and where we work.

According to a study published June 1 by the Federal Reserve Bank of New York, remote work explains approximately 64% of the rise in unemployment among young college graduates since the pandemic. The finding challenges the popular assumption that generative AI tools like ChatGPT are the primary threat to early-career employment.

The Numbers Behind the Trend

The data is striking. The unemployment rate for college graduates under 29 rose 20% — from an average of 3.1% in 2017-2019 to 3.7% in 2022-2025. For graduates aged 22 to 27, unemployment hit 5.8% in 2025, the highest outside the pandemic since 2012.

Yet during the same period, the unemployment rate for more experienced college graduates — those aged 29 and older — actually declined slightly, from 1.9% to 1.8%. This divergence is the puzzle the researchers set out to solve.

As NPR reported, the researchers compared unemployment rates across “remotable” occupations — jobs that can be performed from home, such as software engineering — and “non-remotable” occupations, such as mechanical engineering and nursing. The results were clear: the age gap in unemployment widened significantly in remotable jobs and returned to baseline in non-remotable ones.

The Feedback Deficit

To understand why, the research team — Natalia Emanuel of the New York Fed, Emma Harrington of the University of Virginia, and Amanda Pallais of Harvard — analyzed proprietary data from an unnamed Fortune 500 tech company.

What they found was a stark “proximity penalty.” Software engineers received about 20% more feedback when sitting near colleagues than when working at a distance — a pattern that existed even before the pandemic. After the pandemic shifted everyone to remote work, feedback overall plummeted, and younger workers bore the brunt.

“And that really hit young workers much harder,” Harrington told NPR. “It was these people who had the most to learn that really saw this deficit in feedback.”

The hiring data from the firm told a parallel story. When offices closed during the pandemic, the company hired fewer inexperienced workers and more experienced ones. When it implemented a “pretty aggressive” return-to-office policy, it resumed hiring new graduates. But even after reopening, the firm continued to favor experienced workers for positions on distributed teams.

AI: A Secondary Factor

The study’s timing provides a natural experiment. The rise in youth unemployment predates the rapid diffusion of generative AI tools. When the researchers controlled for occupations’ exposure to AI, remote work remained the dominant factor.

Researchers at the London School of Economics reached a similar conclusion in a working paper examining new hires across the U.S., the U.K., Canada, and Australia — finding that remote work has a clearer impact on early-career hiring than AI.

However, the authors caution that this could change. “It’s always hard to make guesses about what’s going to happen with generative AI,” Harrington said. “It’s certainly possible that this story could really change over the next few years.”

The Scarring Effect

The stakes are high. The NY Fed report warns that entering the labor market during a slack period can have lasting “scarring” effects. Research has consistently found that individuals who begin their job searches in weaker markets tend to have lower earnings and slower career progression relative to peers who started in better conditions.

“Remote work has weakened incentives to hire young workers by impeding on-the-job training,” the researchers wrote. “Employers may not want to hire fresh graduates onto distributed teams because it is more difficult to teach them the requisite skills from afar.”

What This Means Going Forward

The findings carry significant implications for companies, policymakers, and recent graduates alike. For businesses, they suggest that investments in structured remote mentorship programs and virtual onboarding could help bridge the gap. For policymakers, the documented scarring effects of early-career unemployment argue for targeted interventions to help young graduates secure quality first jobs.

Some industry leaders offer a contrasting view. Bill McDermott, CEO of ServiceNow, predicted in March 2026 that AI could push unemployment for new college graduates into the mid-30s within a few years. And Stanford economists have published research finding that in fields where generative AI can automate work, it is indeed harder for young workers to land jobs.

But for now, the evidence points to a more mundane — and more structural — culprit. The shift to remote work, a change that began as a pandemic necessity and has persisted as a permanent feature of the modern workplace, is reshaping the career ladder from the bottom up. And for the class of 2026, that missing rung is making the climb steeper than it has been in years.