Business
Stanford Warns AI Threatens Entry-Level Jobs
A Stanford working paper shows early-career workers in AI-exposed roles saw a 16% employment decline after generative AI spread, signaling a hidden labor market shift.
Photo: Michelangelo Buonarroti / Pexels
A working paper from the Stanford Digital Economy Lab, released in November 2025, indicates a troubling trend in early-career hiring. Workers aged 22 to 25 in the most AI-exposed occupations experienced a 16% relative decline in employment after the spread of generative AI, even after controlling for other factors that might affect firms’ employment decisions. This decline is not observed in more experienced workers or in entry-level jobs with low AI exposure. The concern is specific to early-career jobs that are exposed to AI, such as software developers, customer service representatives, computer programmers, and information systems managers. The paper suggests that firms may be using AI to substitute for the junior tasks through which people traditionally gain their first foothold. This shift is part of a broader trend, with the Federal Reserve Bank of New York reporting that in the fourth quarter of 2025, the unemployment rate for recent college graduates rose to 5.6%, while the underemployment rate reached 42.5%, its highest level since the covid pandemic. *Source: [mittr](https://www.technologyreview.com/2026/05/26/1137865/its-time-to-address-the-looming-crisis-in-entry-level-work/)*
Key points
- Workers aged 22 to 25 in the most AI-exposed occupations experienced a 16% relative decline in employment after the spread of generative AI.
- More experienced workers in those same occupations did not suffer the same decline.
- Employment is also not declining in the entry-level jobs with low AI exposure.
- The concern is specific to early-career jobs that are exposed to AI.
- The Federal Reserve Bank of New York reported that in the fourth quarter of 2025, the unemployment rate for recent college graduates rose to 5.6%, while the underemployment rate reached 42.5%.
- The layer of work AI handles well—translating a specification into routine code, reproducing standard patterns, debugging predictable errors—is precisely the layer that 'learn to code' programs were built around.
- The most effective AI-augmented senior workforce of the late 2030s will be drawn overwhelmingly from the junior cohort of today.