What happened
Fortune reports that sociologist Jeffrey C. Dixon commissioned a YouGov online survey of 1,250 employed U.S. workers between July 30 and Aug. 4, 2026, to examine how artificial intelligence has affected their jobs. About 3% said they had lost a job because of AI since 2023; roughly 6% said they had gained a job that did not previously exist before AI adoption; and approximately 9% said they had received an AI-related promotion or advancement.
Fortune reports that Jeffrey C. Dixon, a sociologist at the College of the Holy Cross, commissioned a 25-question YouGov online survey of 1,250 employed U.S. adults. The survey ran from July 30 through Aug. 4, 2026, and asked about workers’ current AI use, their expectations for future job opportunities, and job changes they believed were connected to AI since 2023. Fortune describes the research as a study in progress whose results have not been peer reviewed.
According to Fortune, approximately 3% of respondents said they had lost a job because of AI, including automation of tasks in their previous job. Roughly 6% said they had landed a job that did not exist before AI adoption, with examples including AI data-labeling and AI compliance work. About 9% said they had received a promotion or advancement they believed was related to AI skills. The source says approximately 95% answered no to the job-loss question, 90% answered no to the new-job question, and 88% answered no to the promotion question. Between 3% and 4% were unsure about each type of change, and rounding may make totals exceed 100%.
Fortune reports that the survey was designed to reflect employed U.S. adults across characteristics including age, gender, race, education and other factors. The article also emphasizes that participants chose to join YouGov’s online panel rather than being randomly selected, which may make their responses differ from those of the broader workforce. Because every respondent was employed when surveyed, the results do not capture people who were jobless after an AI-related layoff or for another reason. The source does not independently verify respondents’ explanations for why their jobs changed.
Why it matters
The findings offer a worker-reported snapshot that is less dramatic than many forecasts about AI-driven employment disruption. They suggest limited reported job loss and some reported career gains so far, while also showing why the labor effects of AI remain difficult to measure and attribute.
Fortune presents the results as evidence that AI has not yet produced substantial reported job losses or gains among the surveyed workers, despite strong predictions from both AI pessimists and optimists. The central finding is not that AI has had no effect, but that relatively few workers in this sample reported a direct AI-related job change since 2023. That distinction matters because broad claims about labor-market transformation can obscure the difference between task changes, job changes and workers’ perceptions of causation.
The reported 9% promotion or advancement figure is notable because it is higher than the reported job-loss and AI-created-job figures. Fortune says AI skills may be helping some workers move up the career ladder, but the article does not establish that AI caused those promotions or that the advantage will persist. It also does not identify which occupations, industries, demographic groups or kinds of AI use were associated with advancement. Those unknowns limit what can be inferred about who benefits from workplace adoption.
The findings may be useful to workers, employers and policymakers because they include reported job gains and promotions alongside job losses rather than measuring displacement alone. At the same time, the survey is a snapshot of employed people’s experiences, not a complete accounting of changes across the labor market. Fortune notes that job changes are generally uncommon, so the small percentages should be interpreted in that context. The source also does not provide independent confirmation from administrative employment records, employers or a peer-reviewed analysis.
What to watch next
The survey’s results should be treated as preliminary. Fortune reports that the underlying study has not been peer reviewed, respondents came from an opt-in online panel, all participants were employed, and workers may not know whether AI caused or contributed to a job change. Future results will need to clarify which workers are advancing, how AI is being used at work, and whether current AI skills retain value as more autonomous systems spread.
Fortune reports that the survey began in July 2026, when the official unemployment rate was 4.1%, while the workforce was shrinking and AI was reportedly cited as a leading reason for many layoffs. The article does not provide enough evidence to determine how those broader conditions relate to the survey responses. Future reporting should distinguish AI-attributed layoffs from ordinary restructuring, economic weakness, outsourcing and other causes, rather than treating every technology-related change as displacement.
The next questions identified by Fortune concern distribution and durability: who is receiving AI-related promotions, whether workers are using AI in ways that improve their careers, and how workers expect AI to affect future opportunities. The source also warns that agentic AI, which can act more autonomously and automate a wider range of work, could reduce the value of some current AI skills. No evidence in this report shows whether that shift is already occurring or how quickly it might affect particular occupations.
Readers should watch for the completed study and for research using more representative sampling or employment records. The current report does not establish a national rate of AI-caused job loss, a forecast of future displacement, or a causal relationship between AI use and promotion. Fortune’s account provides a concrete early measurement, but its conclusions remain constrained by self-reported answers, the opt-in panel, the exclusion of people who were not employed, and the study’s lack of peer review.