A new working paper from Munich's CESifo research institute finds no evidence that AI has significantly reduced hiring of recent college graduates. Analyzing US Census Current Population Survey microdata, researchers Robert Fairlie and Jane Wu looked at unemployment trends for Bachelor's degree holders aged 22 to 25 from 2022 through summer 2026. The summer 2026 unemployment rate of 7.3 percent fell comfortably within the range of previous years, and statistical tests comparing recent graduates with older graduates and non-college graduates showed no meaningful differences.
The findings directly contradict a Stanford study released earlier that used ADP payroll data and found entry-level employment in AI-impacted occupations lagging behind other fields. The CESifo authors suggest the discrepancy may stem from methodology: ADP data tracks the supply of jobs in certain fields, while Census unemployment data also accounts for demand for those jobs. They argue their results serve as a "useful first test" that AI has not yet caused widespread displacement of new graduates.
The paper does note signs that AI adoption is accelerating, including a sharp increase in firms replacing employee tasks with AI and rising AI spending per worker. Venture capitalist Marc Andreessen and BlackRock CEO Larry Fink have both predicted that 2026 graduates could face unusual job market challenges. But the data so far shows a remarkably normal summer for this cohort, with the researchers warning that "if the intensity of AI use in the workplace continues to increase, the graduating classes of 2027 and later might be more affected than the class of 2026."