St. Louis Fed Finds AI Behind One Third of the 18 to 24 Unemployment Jump in 2026

Employment prospects for the youngest workers in America have quietly cratered, and new research from the Federal Reserve Bank of St. Louis explains why. As reported by CIO Dive, the bank's economists William Rodgers III and Alice Kassens found that between April 2023, when the U.S. labor market was at its strongest, and December 2025, the employment rate among 18- to 24-year-olds fell by more than two percentage points. At Metaintro, we track where hiring is actually happening across roughly 50 million job postings, and this study lines up with what the live data has been screaming for months. The entry door narrowed, and it narrowed for the young specifically.

What did the St. Louis Fed actually find?

The headline number is stark. In a June 30 research post, Rodgers and Kassens reported that from April 2023 to December 2025 the employment-to-population ratio for 18- to 24-year-olds moved by -2.24 percentage points, while their unemployment rate climbed 3.51 percentage points and labor force participation slipped 0.35 percentage points. The unemployment jump is the tell. If young people had simply given up and stopped looking, participation would have collapsed. Instead, as the authors put it, the deterioration appeared primarily as higher unemployment rather than as labor force exits, indicating that younger workers were still searching for jobs but with fewer opportunities available.

Now compare that to everyone older. For workers aged 25 to 64, the same three measures barely moved over the identical window, with an employment-to-population change of 0.15 percentage points, an unemployment change of 0.63 percentage points, and a participation change of 0.68 percentage points. Once you account for overall labor demand, there was no comparable slide for prime-age workers, whose employment outcomes remained largely stable. That contrast is the whole story in one line. The pain is not spread evenly across the workforce. It is concentrated at the exact moment a person is trying to get in.

How much of this is really about AI?

This is where the study earns its keep, because it refuses the easy narrative. Yes, AI is part of it. The economists calculated that roughly one-third of the increase in the unemployment rate for 18- to 24-year-olds was attributable to rising demand for the skills needed for AI jobs. Nationally, growing demand for AI-related work accounted for about forty-five percent of the decline in their employment-to-population ratio and roughly one-third of the rise in their unemployment rate. In the Eighth Federal Reserve District, which covers all of Arkansas, most of Missouri, and parts of Illinois, Indiana, Kentucky, Mississippi and Tennessee, AI job demand accounted for nearly half of the increase in young workers' unemployment rate.

But here is the part that should reframe your job search. The researchers were emphatic that the business cycle, not the robots, did most of the work. A broad decline in overall job openings accounted for the largest share of young workers' deteriorating outcomes across all three measures. AI mattered, the authors wrote, but in a narrow, early and age-specific way. AI is not eliminating jobs across the economy. Instead, it is raising the bar for young workers trying to secure their first foothold in the labor market, an area in which employers feel less of a need to advertise openings, make offers and hire. Read that twice, because it changes the advice. If the primary problem were AI automating roles out of existence, there would be nothing to do but retreat. Because the primary problem is a frozen hiring market that has stopped advertising entry roles, the move is to get in front of employers who are hiring quietly rather than posting publicly.

The decomposition also shows how the two forces stack unevenly across groups. For new entrants with no more than a high school diploma, falling job openings drove the bulk of the unemployment increase while the AI-related share was comparatively small. For recent college graduates, AI-related demand carried noticeably more weight, which fits the intuition that the roles most reshaped by generative tools sit in the degree-holding, white-collar lane rather than in service and trade work. That split is worth internalizing when you plan your next move. If your target field is one where postings increasingly list AI competencies, the bar is rising on you specifically, and closing that gap is the single highest-leverage thing you can do before your next application.

Why does a hiring slowdown hit young people first?

The St. Louis Fed calls the current environment a low-hire, low-fire economy, and the phrase deserves unpacking because it explains the age gap perfectly. In this kind of market, companies hold on to the workers they already have and scale back on new hiring. Layoffs stay low, which is why the unemployment rate for prime-age workers looks calm. But hiring also stays low, and hiring is the only channel through which a 22-year-old with no track record can enter. When the front door closes, the people standing outside it are, by definition, the ones who were about to walk in for the first time.

This is why the authors stress that hiring slowdowns show up first and most clearly among young and inexperienced workers. Experienced workers are insulated by inertia, since replacing them is expensive and risky, so firms keep them. New entrants have no such protection. They are the marginal hire, the one a nervous manager can defer for another quarter. For your career, the implication is that generic patience is a bad strategy in a low-hire market. Waiting for the flood of postings to return can cost you a year. The workers who break in are the ones who target the specific companies that are still adding headcount and who reach hiring managers directly instead of feeding applications into a queue that is barely moving.

Are college graduates being hit even harder?

Yes, and the numbers on recent graduates are worse than the overall youth figures. In the St. Louis Fed's decomposition, recent college graduates saw their employment-to-population ratio move by -3.23 percentage points, with their unemployment rate up 1.98 percentage points and participation down 1.58 percentage points over the April 2023 to December 2025 window. For this group specifically, demand for AI jobs did appear to be displacing some new entrants, though falling job openings still explained more of the increase in their unemployment than any other single factor.

That said, the picture is not a wipeout. Even at the end of the period, recent graduates still held relatively high employment-to-population ratios, at 79% nationally and 82% in Eighth District states in the first quarter of 2026. The risk, the researchers noted, comes from a lack of experience and a concentration in entry-level roles precisely as job requirements shift toward AI tasks. The practical read for a new graduate is that a degree still carries weight, but it no longer clears the bar on its own. The graduates who are landing offers are the ones who can point to a concrete project, a portfolio, or demonstrated fluency with the AI tools their target field now assumes, rather than relying on the credential to speak for them.

What role is remote work playing in all this?

There is a second, quieter force in the data, and it involves where the work gets done. Researchers from the Federal Reserve Bank of New York found that remote work may be a factor, particularly among recent college graduates. In their analysis, younger workers' unemployment rate went up one percentage point in jobs that can be done remotely, while older workers' unemployment rate in those same remote-capable roles actually declined slightly. The direction is counterintuitive until you think about training. In a fully remote role, an inexperienced hire is harder to bring up to speed.

Tellingly, younger workers fared better in occupations that could not be performed remotely. The hiring patterns of one Fortune 500 firm suggested the company was willing to teach junior workers when proximity was feasible but shied away from employing inexperienced workers if distance created barriers to training and development. For a job seeker, this is an unexpectedly useful signal. If you are early in your career and struggling to get remote roles, do not read it as a personal failing. It may be structural. Roles that involve in-person work, or hybrid roles where a manager can actually train you at a desk nearby, may be a faster on-ramp than chasing the fully remote listings that draw thousands of applicants and reward proven independence.

How did the researchers measure the AI effect?

It is worth understanding the method, because it determines how much to trust the conclusion. Rodgers and Kassens combined three datasets. For labor force outcomes they used individual-level data from the Current Population Survey, the monthly survey sponsored jointly by the U.S. Census Bureau and the U.S. Bureau of Labor Statistics. To capture general labor demand they used job openings from the Job Openings and Labor Turnover Survey, the BLS series that tracks how many roles employers are trying to fill. And to isolate AI specifically, they turned to detailed job-posting data from Lightcast, classifying a posting as an AI job if it required skills from one of ten AI skill clusters in areas such as generative AI, machine learning and neural networks.

The key nuance, which the authors flag directly, is what this measures. The posting data captures how job content is changing, meaning the skills and tasks employers list, not whether firms are using AI to automate jobs or screen applicants. In plain terms, the study is not saying an algorithm rejected you. It is saying the job description itself now expects AI competence that entry-level candidates often do not yet have. That distinction matters for your plan, because a rising bar in the job description is something you can meet with a few weeks of deliberate skill-building. An automated rejection would be far harder to route around.

Does other research back this up?

It does, and the corroboration is what makes the story credible rather than a single provocative chart. The St. Louis Fed's work sits alongside a 2025 study out of Stanford University, which found that since late 2022, when OpenAI launched ChatGPT and generative AI became widespread, the employment of early-career workers in the most AI-exposed fields, such as software engineering and customer service, has declined significantly. Two independent research teams, using different data, are pointing at the same demographic and the same timeframe.

What none of this research supports is panic. The consistent finding across the St. Louis Fed and the corroborating work is that the effect is concentrated, early-stage, and specific to the point of labor market entry. The economists themselves frame their next post around how disconnected young adults, those neither in school nor working, spend their time, which underscores that the concern is about access to that first rung, not the disappearance of careers wholesale. For a job seeker, the actionable version is simple. The market has raised the entry bar and thinned the number of open doors at the same time. You cannot control the business cycle, but you can control whether you show up to an employer already fluent in the tools their postings now demand, and whether you find the quiet openings that never get advertised.

Related Articles

• Low-Hire, Low-Fire Job Market and What It Means for Your Search in 2026
• Entry-Level Roles Are Disappearing in the 2026 Job Market
• Entry-Level Grads Taking Any Paycheck in a Brutal 2026 Market
• Teen Employment Decline in June 2026 and Young Job Seekers
• The Gen Z Job Market Is Not AI's Fault
• A Shortage of Openings, Not AI, Is Behind the Gen Z Job Hunt in 2026
• Colleges Keep Producing Graduates the 2026 Job Market Will Not Hire
• How to Stop AI From Becoming the Enemy of Younger Workers
• The US Labor Market Is Splitting Between AI Office Jobs and Construction Hiring
• Which College Degrees Pay Off in 2026
• Young Adults, Entry-Level Jobs, and the 2026 Economy

People Also Asked

Q: How much did employment fall for 18 to 24 year olds?

A: According to the Federal Reserve Bank of St. Louis, the employment rate for workers aged 18 to 24 fell by more than two percentage points between April 2023 and December 2025. In the detailed decomposition, their employment-to-population ratio moved by 2.24 percentage points and their unemployment rate rose by 3.51 percentage points, while workers aged 25 to 64 stayed largely stable over the same window.

Q: Is AI the main reason young people cannot find jobs?

A: No. The St. Louis Fed found that AI demand explained roughly one-third of the rise in the 18 to 24 unemployment rate and about forty-five percent of the decline in their employment-to-population ratio, but a broad decline in job openings was the largest single force. The economists described AI's effect as narrow, early and age-specific, and stressed it is raising the bar for entry rather than eliminating jobs across the economy.

Q: What should a recent graduate do about the tighter entry-level market?

A: Recent graduates still held relatively high employment-to-population ratios, at 79% nationally and 82% in Eighth District states in early 2026, so the situation is tighter, not hopeless. The practical moves are to build demonstrable AI fluency in the tools your target field now expects, favor in-person or hybrid roles where employers are more willing to train juniors, and reach hiring managers directly since many entry openings are no longer advertised.

Stay ahead of the market by seeing where employers are actually hiring before the postings dry up. Metaintro reads the live job data across roughly 50 million postings so you can target the companies still adding entry-level headcount instead of guessing. Create a free Metaintro account and put the labor market's real signals to work on your job search.