At companies that have adopted generative AI, entry-level hiring has fallen by roughly 80% per quarter since 2023, according to a new working paper from Harvard University. AI is reducing demand for junior workers across industries, and the effects are already visible in entry-level hiring.
The research, by Harvard economists Seyed Hosseini and Guy Lichtinger, analyzed résumé and job posting data covering 66 million workers across more than 280,000 U.S. firms between 2015 and 2025. Generative AI is disproportionately reducing demand for workers at the bottom of the career ladder while senior employment at the same firms continues to grow. Hosseini and Lichtinger call it “seniority-biased technological change.”
The pattern holds across industries and isn’t driven primarily by layoffs. Companies are pulling back on entry-level hiring, and the reasons are more complicated than AI simply replacing jobs.
Entry-Level Hiring Falls As AI Automates Routine Work
In many white-collar jobs, entry-level workers spend their early years handling intellectually routine tasks, including debugging code, reviewing legal documents, drafting communications and entering data. These are exactly the kinds of tasks generative AI handles best.
At companies that actively integrated AI into their workflows, entry-level employment fell roughly 9% within six quarters of adoption relative to companies that didn’t. At the same time, senior employment at those firms continued to grow.
The decline reflects the nature of entry-level work itself. Generative AI is particularly effective at codified, checkable tasks, the kind learned from textbooks and training data rather than accumulated on the job. The tasks that define early careers fall squarely into that category. For companies building AI into daily workflows, that means fewer entry-level workers.
Entry-Level Hiring Drops As Firms Bet On Future AI Gains
The 80% decline in entry-level hiring didn’t happen because automation had already replaced those workers. It happened because companies expected it would.
Following the release of ChatGPT in late 2022, mentions of AI in U.S. firms’ earnings calls tripled by mid-2023, according to the researchers. Companies began pulling back on entry-level hiring almost immediately, adjusting for automation they anticipated rather than automation that had already arrived.
The decline is driven primarily by slower hiring, not increased layoffs. Separation rates for entry-level workers fell at adopting firms. Workers who left simply weren’t replaced. If companies overestimate how quickly AI will automate entry-level work, some of these cuts may ultimately prove premature.
Entry-Level Hiring Shifts Toward More Experienced Candidates
At AI-adopting firms, senior employment continued to grow throughout the same period that entry-level hiring contracted. Within those firms, entry-level workers in occupations most exposed to generative AI saw a 7% relative decline compared to workers in less exposed roles. Senior workers in those same occupations saw no comparable drop.
A Stanford analysis of ADP payroll records found a 16% relative decline in employment among early-career workers in the most AI-exposed occupations since late 2022. The declines were most pronounced in software development and customer service, two fields that have historically absorbed large numbers of entry-level workers and served as reliable on-ramps to white-collar careers.
The pattern reflects a broader shift in how companies value experience. Generative AI is making senior workers more productive while reducing the need for the routine work that entry-level workers typically perform. Companies aren’t just hiring fewer entry-level candidates. They’re concentrating demand at the top of the experience ladder.
Entry-Level Hiring Stalls As Companies Redesign Jobs Around AI
The shift isn’t only about headcount. At AI-adopting firms, tasks most exposed to generative AI are being explicitly removed from entry-level job postings.
Data from the Anthropic Economic Index, which classifies real-world interactions with large language models by task type, found that AI replaces work outright in more than 50% of interactions in entry-level occupations. For senior occupations, that figure drops to 40%.
The work that traditionally defined entry-level roles isn’t just being automated. It’s disappearing from job descriptions altogether.
You might expect companies to backfill that lost work by moving entry-level workers into more complex responsibilities. Hosseini and Lichtinger looked for exactly that and found no evidence of it. Companies aren’t handing entry-level workers more sophisticated work as AI absorbs routine tasks. They’re simply eliminating the roles.
Entry-Level Hiring Weakens When Training Pipelines Lag Behind AI
Entry-level jobs have always been about more than filling roles. They’re how companies develop the senior talent they’ll need years from now. When companies stop hiring entry-level workers, they weaken the pipeline that produces experienced candidates tomorrow.
There are signs that some companies recognize the risk. Eighty-eight percent of chief human resources officers say AI is making early-career talent role-ready faster, according to a survey by SAP and Wakefield Research. That’s an argument for redesigning entry-level roles around AI, not eliminating them. Companies that figure out how to bring entry-level workers in and develop them alongside AI tools will have a meaningful advantage over those that simply stop hiring.
What Entry-Level Hiring Says About Your Talent Strategy
The case for entry-level hiring isn’t just about fairness to new graduates. It’s about organizational continuity. Senior workers don’t appear from nowhere. They develop through the entry-level roles companies are now cutting, and the skills they build early in their careers compound over time in ways AI can’t easily replicate.
The timing compounds the risk. Companies making hiring cuts today are doing so at a time when AI tools are still developing, and the automation gains they’re anticipating haven’t fully materialized. If the early evidence holds, the costs will fall unevenly on workers who never got the chance to start and on organizations that later find themselves without experienced talent.
Hosseini and Lichtinger are careful not to overstate their conclusions. The data covers a relatively short window, and longer-term adjustments in how companies train and develop workers could still change the picture. But the early evidence is difficult to ignore. Generative AI adoption, they write, “may be shifting work away from entry-level tasks, potentially narrowing the bottom rungs of internal career ladders.” For companies still treating entry-level hiring as optional, that’s the risk worth addressing now rather than later.