New Study Finds No Evidence AI Is Harming Recent Graduate Employment

A new working paper from economics researchers at Munich’s CESifo challenges the idea that artificial intelligence is hurting entry-level employment. The researchers argue that there is no evidence of any significant, widespread displacement or reduction in hiring of recent college graduates in absolute or relative levels.

In their study titled The Early Impacts of AI on Employment Among Recent College Graduates, authors Robert Fairlie and Jane Wu focused on new degree holders because labor demand shifts often appear first through hiring cuts. They noted that as AI handles standardized tasks in entry-level office jobs, firms might reduce new hiring rather than lay off experienced workers.

Context for the study includes reasons why 2026 graduates might face higher risks than previous classes. The researchers pointed to Census survey data showing a sharp increase in firms replacing tasks with AI, along with growth in AI spending per employee and ChatGPT Enterprise token use over the past year.

Prominent figures also expressed concern prior to the study. Venture capitalist Marc Andreessen stated earlier this year that AI was not actually good enough until December 2025 to do the jobs currently being cut. BlackRock CEO Larry Fink warned in March that rapid AI changes could cause the highest new graduate unemployment rate in years.

To check these worries, the CESifo researchers analyzed microdata from the US Census Current Population Survey. They focused on Bachelor’s degree recipients aged 22 to 25 who were not pursuing higher degrees. Because summer unemployment naturally spikes as graduates enter the market, the team examined seasonal and year-over-year trends starting in 2022.

The summer 2026 unemployment rate for young graduates stood at 7.3 percent. This figure fell within the range of previous years, spanning from 6.3 percent in 2022 to 7.8 percent in 2024. Results remained similarly unremarkable when including graduates who wanted a job without actively looking.

The team used statistical tests to compare recent graduates with non-college peers in the same age bracket and with older college graduates aged 30 to 49. They also categorized employment by potential AI exposure using a 2023 study on job roles suited for AI systems.

Most comparisons showed that trend differences between groups from 2022 to 2026 lacked statistical significance. The authors concluded that summer 2026 unemployment for recent graduates was not unusually high compared to earlier summers across the tested groups.

This outcome differs from a recent Stanford University study that found lagging entry-level employment in AI-impacted fields. The Stanford research relied on payroll data from HR firm ADP, which covers a different cross-section of the economy and tracks job supply rather than aggregate job demand.

The CESifo authors noted that current trends do not guarantee future performance. They warned that if workplace AI intensity keeps growing, graduating classes in 2027 and later could face heavier impacts, requiring additional years of data to monitor emerging effects.