Jeffrey C. Dixon, a professor of sociology at the College of the Holy Cross, commissioned a YouGov survey to ask US workers how artificial intelligence has affected their employment since 2023.
Survey methodology
The online questionnaire was completed by 1,250 employed adults between 30 July and 4 August 2026. Respondents were selected to mirror the broader US labour force in terms of age, gender, race, education and other demographics. The 25-question instrument included three items directly probing AI-related job changes:
- Whether the respondent had lost a job because AI automated their previous tasks.
- Whether they had taken a newly created role, such as an AI data labeler or compliance specialist.
- Whether they had received a promotion or advancement they attributed to AI skills.
For reference, the official unemployment rate at the time of the survey was 4.1% according to the U.S. Bureau of Labor Statistics.
Key findings
Only about three per cent of those surveyed said they had lost a job due to AI, while roughly six per cent reported securing a position that did not exist before the rise of AI. Around nine per cent indicated they had earned a promotion linked to AI expertise. The majority, 95% on the first question, 90% on the second and 88% on the third, answered "no". A small share (3-4%) were unsure of any AI-related change.
Why it matters
These numbers suggest that, four years after the launch of ChatGPT in November 2022, AI has not yet triggered the sweeping labour market upheaval predicted by some analysts. The modest incidence of job loss contrasts with the more noticeable, albeit limited, opportunities for advancement, indicating that AI skills may be becoming a modest career lever for a minority of workers.
Experts warn that newer "agentic" AI systems, capable of greater autonomous action, could alter the value of current AI expertise, potentially reshaping the balance between job displacement and creation.
What happens next
Dixon plans to extend the research to explore who is most likely to receive AI-related promotions and how workers are integrating AI tools into daily tasks. Further questions include how expectations of AI's future impact vary across occupations and demographic groups.
While the survey provides a useful snapshot, its reliance on an online panel means the findings may not fully represent the entire US workforce, especially those who are currently unemployed. Nonetheless, the data offers a counterpoint to more alarmist forecasts and highlights the need for ongoing monitoring as AI technologies evolve.