AI is impacting jobs, but not all kinds of roles are equally being impacted.
A groundbreaking study from Harvard University reveals that the adoption of generative AI is leading to a significant decline in junior-level positions while senior roles continue to expand. The research, which analyzed a massive dataset of U.S. résumés and job postings, provides some of the first concrete evidence of a “seniority-biased technological change” in the age of AI.
The paper, authored by Seyed M. Hosseini and Guy Lichtinger, examined the employment dynamics within nearly 285,000 U.S. firms between 2015 and 2025. Their findings indicate a stark divergence in employment trends that began in early 2023, coinciding with the widespread proliferation of generative AI tools.
Methodology: Tracking AI Adoption and its Impact
To measure the real-world impact of AI, the Harvard researchers developed a novel method to identify firms that were actively integrating the technology. They analyzed millions of job postings, flagging those seeking “AI integrator” roles—positions dedicated to implementing and operating AI systems within a company. This approach allowed them to move beyond theoretical exposure to AI and focus on actual adoption.
The study utilized a massive dataset from Revelio Labs, covering nearly 62 million workers and their employment histories. This data, which includes standardized seniority levels for each position, enabled the researchers to track the changes in the number of junior (Entry/Junior) and senior (Associate and above) employees within firms over time.
Key Findings: A Widening Gap Between Junior and Senior Employment
The study’s central finding is a sharp, relative decline in junior employment at firms that have adopted generative AI. Key observations from the research include:
- Diverging Employment Trends: While junior and senior employment grew at similar rates from 2015 to mid-2022, a significant gap emerged thereafter. Junior employment flattened and then began to decline in early 2023, just as senior employment continued its upward trajectory.
- Impact of AI Adoption: In firms that adopted AI, junior employment fell by 7.7 percent relative to non-adopting firms within six quarters, starting from the first quarter of 2023. In contrast, senior employment in these same firms continued to grow, a trend that had been observed even before the widespread adoption of AI.
- Hiring, Not Firing: The reduction in junior positions is primarily driven by a slowdown in hiring rather than an increase in separations or layoffs. AI-adopting firms significantly reduced their hiring of junior workers but also saw a modest decrease in separations for this group.
- Increased Promotions: Interestingly, the study found that promotions for existing junior employees into more senior roles increased in firms that adopted AI. This suggests that while fewer new juniors are being hired, those already in the company may have more opportunities for advancement.
- Sector-Wide Impact: The decline in junior hiring was observed across various sectors, with the most significant drop in wholesale and retail trade, where AI-adopting firms hired approximately 40% fewer juniors.
- A “U-Shaped” Effect on Graduates: The impact on junior employees also varied by their educational background. Graduates from mid-tier universities experienced the most substantial declines in employment. Those from elite and lower-tier institutions were less affected, with the smallest impact seen among graduates from the lowest-tiered schools.
The Bigger Picture: Substantiating a Shifting Labor Market
The findings of the Harvard study are consistent with a growing body of evidence and anecdotal reports from the tech and business sectors. A recent Stanford study also found that since late 2022, employment for young, entry-level workers in occupations most exposed to AI has seen a significant relative decline. This trend is particularly noticeable in fields like software engineering and customer service.
The underlying reason for this shift appears to be the nature of entry-level work. Many junior roles in white-collar professions involve routine, cognitively demanding tasks like debugging code, reviewing documents, or drafting standard communications—tasks that are highly susceptible to automation by generative AI. As one executive at a recruiting firm noted, the work of young graduates, once in high demand, is now a “home run’ for AI.” This sentiment is echoed by concerns that AI is eroding the “bottom rungs” of the career ladder, making it more difficult for new entrants to gain a foothold and develop their skills.
The implications of this seniority-biased shift are far-reaching. As the Harvard paper concludes, if AI disproportionately affects junior positions, it could have “lasting consequences for the college wage premium, upward mobility, and income disparities.” With a significant portion of lifetime wage growth stemming from early-career advancements, the narrowing of entry-level opportunities could reshape career trajectories for a generation. Firms may need to rethink their talent development strategies, potentially relying more on experienced hires and accelerating internal promotions. For individuals entering the workforce, the ability to work alongside AI and focus on less automatable, higher-order skills will be more critical than ever.