AI is hitting entry-level jobs hardest, Stanford study finds
Stanford researchers found that employment for workers aged 22-25 in the most AI-exposed occupations is now 19% below peers in less exposed fields, up from 13% the prior year The employment gap is driven primarily by lower hiring rates for entry-level workers, not increased firings or attrition AI's impact is highly differentiated: "automative" uses (fully replacing human work) correlate with declining entry-level employment, while "augmentative" uses show flat or rising employment Jobs relying
Analysis
TL;DR
- Stanford researchers found that employment for workers aged 22-25 in the most AI-exposed occupations is now 19% below peers in less exposed fields, up from 13% the prior year
- The employment gap is driven primarily by lower hiring rates for entry-level workers, not increased firings or attrition
- AI's impact is highly differentiated: "automative" uses (fully replacing human work) correlate with declining entry-level employment, while "augmentative" uses show flat or rising employment
- Jobs relying on codified knowledge (formal, documented, teachable) are disproportionately affecting younger workers, while tacit knowledge roles (gained through experience) protect mid-career and senior workers
- Higher education appears to serve as a buffer, with occupations having more college graduates showing muted differences between AI-exposed and less-exposed roles
Why It Matters
This research provides some of the first empirical evidence that AI's labor market effects are not uniformly distributed but are instead concentrating on entry-level workers in specific types of occupations. For AI practitioners and policymakers, it underscores that the immediate impact of AI is not a blanket "jobs apocalypse" but a targeted disruption that could reshape career pipelines, particularly for younger entrants into fields like accounting, administration, and other codified-knowledge roles.
Technical Details
- The study uses anonymized, high-frequency payroll data from ADP, a major HR management company, covering a large subsample of the U.S. workforce
- AI exposure was measured using two complementary metrics: a potential labor market impact gauge from prior research and the Anthropic Economic Index, which analyzes actual Claude model usage patterns across occupations
- The Anthropic Economic Index distinguishes between "automative" queries (tasks fully replaceable by AI) and "augmentative" queries (tasks where AI enhances human productivity), providing a nuanced exposure classification
- Codified knowledge was proxied using O*NET's occupational database, specifically the required level of formal education per occupation
- The analysis compares employment trends for workers aged 22-25 between the top 40% most AI-impacted jobs and the bottom 60% least impacted, tracking changes since 2022
Industry Insight
- Organizations should anticipate a tightening entry-level hiring market in AI-exposed, codified-knowledge fields and consider restructuring training programs or investing in AI-augmentation workflows rather than pure automation to preserve career pipelines
- Workers and educators should recognize that roles emphasizing tacit, experience-based knowledge are more resilient to AI displacement at the entry level, suggesting strategic value in developing mentorship-heavy and practice-based skill pathways
- The widening gap between automative and augmentative AI use cases suggests that companies adopting AI as a complement to human workers may face fewer workforce disruption risks than those pursuing full task automation, particularly for junior roles
Disclaimer: The above content is generated by AI and is for reference only.