Not across the U.S. labor market in the period it measured. A Yale Budget Lab analysis published October 1, 2025 found no discernible economy-wide disruption in its labor-market measures during the first 33 months after ChatGPT launched in November 2022. That is a short-run finding—not proof that no worker has been displaced, or that AI will not affect jobs later.
What did Yale’s study find?
The report, Evaluating the Impact of AI on the Labor Market: Current State of Affairs, examined U.S. labor-market data through the latest monthly Current Population Survey release available to its authors in July 2025. The authors concluded that “the picture of AI’s impact on the labor market that emerges from our data is one that largely reflects stability, not major disruption at an economy-wide level.”
That conclusion is about broad patterns in the measured period. An economy-wide measure can remain relatively stable even while particular workers, occupations or employers experience changes. The study does not establish that AI caused—or did not cause—any specific job loss.
How did the researchers measure labor-market change?
The authors tracked changes in the distribution of workers across occupations using monthly Current Population Survey data. Their dissimilarity index compares occupational shares over time; a 12-month moving average helps smooth monthly noise. The researchers compared the period following ChatGPT’s public launch in November 2022 with earlier periods associated with personal-computer popularization (1984–1989), internet adoption (1996–2002), and a 2016–2019 control period.
#1 Best Overall
The index measures shifts in occupational composition, not a tally of jobs eliminated by AI. A change can reflect people switching occupations, entering or leaving employment, or finding work in a different occupation. The index alone cannot identify the reason for the change.
How much did the occupational mix change?
At a comparable point in the time periods, the post-ChatGPT occupational-mix path was about one percentage point above the internet comparison. The report describes that difference as modest, and notes that occupational changes were already occurring before ChatGPT appeared.
Rank #2
Some industries showed larger changes than the overall labor market. The Information sector’s occupational-mix change was around 14% by 32 months, compared with just over 4% at the baseline. Those figures describe a change in the sector’s occupational mix; they are not estimates of AI-caused job losses. The report says shifts in Information, Financial Activities, and Professional and Business Services were underway before ChatGPT, and that the Information sector’s longer-run changes appear characteristic of the sector rather than attributable to one technology.
Does AI exposure mean workers are using AI or losing jobs?
No. The study’s exposure measure uses OpenAI task-level estimates to classify how much work in an occupation could theoretically be affected. The report says worker shares across its relative exposure groups remained broadly stable after ChatGPT’s launch: about 29% low exposure, 46% medium exposure and 18% high exposure. These are exposure categories, not estimates of workers who used AI or lost jobs.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
To examine observed use, the authors also considered data about Claude, an AI tool made by Anthropic. That evidence covers one tool rather than all workplace AI, and the report notes that Claude activity is concentrated in some occupational groups and does not cover all tasks or tools. The authors treat both exposure estimates and single-tool usage data as imperfect proxies, not comprehensive measures of adoption across the economy.
What about recent college graduates?
The report notes a slight recent increase in occupational-mix dissimilarity between recent and older college graduates. It does not establish that AI reduced hiring of young graduates. The authors caution that the relevant CPS samples are small and noisy, and that the observed pattern may have begun before ChatGPT or reflect a broader slowing labor market.
What the study can—and cannot—say
- It can describe: measured U.S. employment patterns and occupational composition during the study window, and how those patterns compare with selected earlier technology periods.
- It cannot establish: that AI caused a particular worker’s job loss, that no displacement occurred, or that AI’s effects will remain limited.
- It does not measure comprehensively: workplace AI adoption across tools, enterprise systems and API use.
The report is observational, and its authors explicitly warn: “Of course, our analysis is not predictive of the future.” The authors say more comprehensive, privacy-protected data from leading AI companies—including enterprise and API use—would help clarify how AI is being adopted at work.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What this means if you are thinking about your job
The study offers a reason not to treat dramatic claims about economy-wide job replacement as an established description of the first 33 months after ChatGPT’s launch. It does not promise that any particular role is safe. For personal planning, distinguish between a task that AI might assist with and evidence that employers are using AI to reduce staffing; this report does not settle that question for an individual occupation, workplace or later period.
Best Value
The headline framing also appeared in ITPro’s October 1, 2025 coverage, “AI isn’t taking anyone’s jobs, finds Yale study – at least not yet”. The Yale Budget Lab report is the primary source for the study’s methods, figures and qualifications.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
