AI Hasn’t Yet Raised Unemployment for 2026 College Graduates, Study Finds

Artificial intelligence has not yet caused an unusual rise in unemployment among recent college graduates, according to a CESifo analysis of U.S. employment data. The findings suggest that the early impact of AI on the 2026 job market remains difficult to detect.

2026 graduate unemployment remains within the normal range

The summer unemployment rate for young college graduates in 2026 was 7.3 percent. That figure falls within the previous years’ range, which ran from 6.3 percent in 2022 to 7.8 percent in 2024.

The results were similarly unremarkable when the analysis included graduates who were not actively looking for work but said they wanted a job in the Current Population Survey. Those people are not officially counted among the unemployed workforce.

Researchers compared college graduates with other workers

To test the reliability of the results, the researchers developed a statistical test comparing recent college graduates with non-college graduates and older college graduates between 30 and 49 years old.

They also categorized jobs by their potential exposure to AI, using a 2023 study that examined which types of work AI systems are best suited to perform.

Across nearly all comparisons, differences in employment trends between the groups were not statistically significant during the 2022–2026 period. The researchers wrote that the data “tells a consistent story that unemployment rates for college graduates in the summer of 2026 were not unusually high compared to previous summers.”

Why the findings differ from a Stanford study

The results appear to conflict with a recent Stanford study that reported almost completely opposite findings. However, the two analyses use different data sources and measure different aspects of the labor market.

The Stanford study relies on payroll data from human resources firm ADP. Although ADP covers a significant portion of the economy, it may not capture every element of the broader census. ADP data also measures the aggregate supply of jobs in different sectors, while the unemployment rates examined by CESifo account for aggregate demand as well.

Even if the supply of jobs in a particular field begins to shrink because of AI, demand for those jobs could change as well.

AI’s long-term effect on jobs is still uncertain

The CESifo researchers describe the summer 2026 unemployment data as a “useful first test” of how much AI is affecting the current U.S. job market. So far, the data suggests that AI has not had a measurable impact on unemployment among recent college graduates.

However, the researchers caution that current trends do not predict future outcomes. If workplace AI adoption continues to increase, graduates in 2027 and beyond may be more affected than those entering the workforce in 2026. They add that many more years of data will be needed to determine whether AI’s impact grows as its use in the workplace deepens.