AI-Exposed Majors Lost 13% of Starting Earnings After ChatGPT

New college graduates in the most AI-exposed college majors saw their initial earnings fall 13 percent after ChatGPT's release in late 2022, according to a U.S. Census Bureau working paper published in September 2026. The drop is comparable in size to the earnings losses recorded by graduates who enter the labor market during a large recession.

The paper, "Graduating into Disruption: Labor Market Outcomes for AI-Exposed College Majors" (Working Paper CES-26-56), was written by Census Bureau researchers Cody Orr, Lee C. Tucker and Lawrence Warren. It uses administrative records on college graduates to track employment, earnings and job-switching outcomes across majors with differing levels of AI exposure since large language models became widely available.

What did the researchers measure?

The study ranks college majors by labor market AI exposure and compares outcomes before and after ChatGPT's introduction in late 2022. In regression-adjusted estimates, the most AI-exposed decile of majors saw two distinct effects begin immediately after the tool's launch:

• The likelihood of initial post-graduation employment fell by five percentage points.
• Full-quarter initial earnings declined by 13 percent.

Employment, earnings and job-switching patterns in the most exposed majors began to diverge from less exposed majors right away, not gradually. The authors read this timing as evidence on the mechanisms behind AI's broader labor market effects.

Where did the earnings losses come from?

Roughly half of the 13 percent earnings decline reflects lower pay within the industry sectors that traditionally employ these graduates. The other half comes from graduates shifting into lower-wage industry mixes — sectors such as restaurants and retail.

That split matters for anyone weighing a first job. The penalty is not only a story of employers in exposed industries paying less; it is also a story of graduates being pushed out of their preferred fields entirely and landing in lower-paying work outside them.

Do the effects fade?

The effects attenuate as graduates move further from labor market entry, the paper finds. But they remain substantial for the most exposed majors. In other words, the damage concentrates at the start of a career — exactly when early salary levels typically anchor future earnings trajectories — and does not fully disappear with time.

For a current student or recent graduate in a highly exposed major, the data translate into a concrete market position: a lower probability of landing a first job in the field trained for, a materially lower starting salary if one is secured, and a real chance of starting out in an unrelated, lower-wage sector.

What are the study's limits?

The findings come from a working paper, not a final peer-reviewed publication, and may be revised. The exposure measure applies to college majors grouped by decile, so it captures differences between broad fields of study rather than outcomes for any individual graduate.

The Census Bureau's Center for Economic Studies released the paper in September 2026 as part of its working paper series; the authors say the administrative-records approach offers a direct read on how the arrival of large language models reshaped outcomes for the newest entrants to the U.S. labor market.