Remote Work, Not AI, Drives Youth Unemployment, Fed Finds

Young college graduates are having a harder time finding work than they did before the pandemic, and the popular explanation has been artificial intelligence. Economists at the Federal Reserve Bank of New York looked at the timing and landed on a different prime suspect. Remote work explains 64% of the recent increase in unemployment among young college graduates, their June 1 analysis concluded, a larger share than they attribute to AI.

The finding reorders a debate that has run for the better part of a year. If the machines were not the main thing keeping new graduates on the bench, then the way companies reorganized work after 2020 was, and that moves the remedy from a problem nobody controls to a management choice employers make every quarter.

How the Fed Arrived at the 64 Percent Figure

The analysis came from three economists: Natalia Emanuel, a research economist at the Federal Reserve Bank of New York; Emma Harrington, an assistant professor of economics at the University of Virginia; and Amanda Pallais, a professor of economics at Harvard. Posted on the bank’s Liberty Street Economics analysis of remote work and youth joblessness, it leans on a simple test. Split college-educated workers into jobs that can be done from home, like software engineering and financial analysis, and jobs that cannot, like nursing and mechanical engineering. Then compare how younger and older workers fared inside each group.

The split is stark. Among college graduates under 29, the unemployment rate climbed from 3.1% in the 2017-19 period to 3.7% in 2022-25, a rise of nearly 20%. Among graduates 29 and older, it barely moved, slipping from 1.9% to 1.8%. A separate Fed gauge tracked on the bank’s recent college graduate labor market dashboard shows joblessness for graduates aged 22 to 27 reaching 5.6% in March, up from 3.6% on the eve of the pandemic. The youngest, least experienced workers absorbed almost all of the damage.

Inside the remotable occupations the picture sharpens further. Young workers in jobs that shifted online saw their unemployment rate rise by nearly 1 percentage point between the two periods, while their older colleagues held steady. Run that contrast across the whole labor market and remote work accounts for 64% of the recent jump in joblessness among young college graduates.

Why AI Became the First Suspect

For most of the past year the blame fell on artificial intelligence (AI, the software now drafting code, memos and spreadsheets that junior staff once handled). The story had obvious appeal, and it had data behind it.

In August 2025, Stanford economists led by Erik Brynjolfsson published findings that landed hard. Drawing on payroll records, the Stanford study on AI and early-career employment reported a 13% relative decline in jobs for early-career workers in the most AI-exposed occupations since generative tools spread. Employment for software developers aged 22 to 25 had fallen close to 20% from its late-2022 peak. Older workers in the same fields held steady or grew. Executives amplified the theme, with Dario Amodei, chief executive of the AI firm Anthropic, warning in 2025 that the technology could wipe out half of entry-level white-collar jobs within a few years.

The New York Fed economists do not dispute that AI is reshaping work. They question the sequence. The uptick in youth unemployment, they note, began before generative AI spread through offices, which makes it hard to pin the early damage on a tool that arrived later. Even after accounting for how exposed a job is to AI, the age gap persists in both remotable and non-remotable work.

That ordering matters for anyone trying to fix the problem. A labor market hurt mainly by automation calls for one response, around reskilling and policy. A market hurt mainly by how teams are arranged calls for something cheaper and faster that sits inside any manager’s reach.

How Distributed Teams Freeze Out Beginners

Why would working from home hit the young hardest? The research points to training. New hires learn by watching, asking and getting quick corrections, and those informal channels thin out when teams are scattered.

• Feedback shrinks. Workers who received in-person coaching produced higher-quality output in the Fortune 500 records the authors examined, and younger staff lost the most when that coaching moved online.
• Mentorship is hard to schedule. Spontaneous over-the-shoulder corrections rarely survive being turned into a calendar invite.
• Hiring math changes. A manager who must train remotely often skips the inexperienced candidate and picks someone who can already work without much guidance.

The authors traced this through company personnel records. During pandemic lockdowns, firms hired fewer inexperienced workers across the board. When offices reopened, hiring of young workers bounced back almost everywhere, except in roles that stayed distributed. The beginner penalty clung to remote positions even after the rest of the market normalized, which is the pattern that drives the headline number.

Where Remote Work and AI Overlap

None of this clears AI. Both forces are real, and they often land on the same desks, because the jobs easiest to do from home are frequently the jobs most exposed to automation.

Software engineering sits in both buckets. So do many analyst and clerical roles. A young coder today may be competing with a chatbot for the simple tickets and competing with a remote-first org chart that never built in a seat for a trainee. Separating the two is exactly what the study set out to do, and its answer was that the larger drag on entry-level hiring shows up even where AI exposure is low. For a fuller picture of how the two forces interact, the Stanford Digital Economy Lab keeps a running account of what economists know about AI and the labor market.

The practical takeaway is about proportion. AI gets the headlines and the venture funding, but by this measure the bigger reason new graduates are stuck on the sidelines traces to a workplace habit that predates the chatbots.

What New Graduates Face Now

For the class entering the workforce this spring, the research carries an awkward implication. The remote and hybrid arrangements many young workers prize may be the same ones slowing their start.

Companies may be reluctant to hire less-experienced workers in distributed work arrangements.

That line, from the authors’ summary, is the crux. The fix it implies needs no new technology and no new law. It needs managers to bring beginners into rooms where they can be coached, and to protect the time senior staff spend doing the coaching.

Whether firms choose to is uncertain. Remote work trims real estate bills and widens the hiring pool, and few companies will reverse it to shield a cohort with little bargaining power. 5.6% unemployment for recent graduates is painful for those counted in it, yet it is not the kind of number that forces a boardroom to act on its own.

If employers decide the cost of a stalled generation outweighs the savings of an empty office, the youngest workers gain ground. If they do not, the next graduating class inherits the same arithmetic.

Frequently Asked Questions

Does Remote Work or AI Cause Youth Unemployment, According to the New York Fed?

Remote work, the bank says. Its June 1 analysis attributes 64% of the recent rise in unemployment among young college graduates to the spread of remote work, a larger share than it links to AI, partly because youth joblessness began climbing before generative tools were widely adopted.

How Much Has Unemployment Risen for Recent College Graduates?

The unemployment rate for recent graduates aged 22 to 27 reached 5.6% in March, up from 3.6% in March 2019. For college graduates under 29 more broadly, it rose from 3.1% in 2017-19 to 3.7% in 2022-25, while the rate for older graduates barely changed.

Why Does Remote Work Hurt Younger Workers More Than Older Ones?

New employees build skills through in-person feedback and mentorship, and those channels weaken on distributed teams. The Fed economists found managers often avoid hiring inexperienced candidates for remote roles because training someone from a distance is harder than coaching them in person.

What Did the Stanford Study Say About AI and Entry-Level Jobs?

Stanford economists led by Erik Brynjolfsson reported in August 2025 a 13% relative decline in employment for early-career workers in the most AI-exposed occupations, with software developers aged 22 to 25 down close to 20% from their late-2022 peak.

Who Wrote the New York Fed Analysis?

Natalia Emanuel of the Federal Reserve Bank of New York, Emma Harrington of the University of Virginia, and Amanda Pallais of Harvard. The analysis was published June 1, 2026 on the bank’s Liberty Street Economics blog.