| Notes |
This series reports the share of total work hours assisted by generative artificial intelligence among employed adults aged 18-64 in Construction and Extraction Occupations. Employed respondents who used generative artificial intelligence for their job last week report on how many workdays they used it (one, more than one, or every workday) and how much time they spent actively using it on those days (less than 15 minutes, 15 minutes to 1 hour, 1 to 4 hours, or more than 4 hours per day). Because both responses are bracketed, they imply a lower and an upper bound on assisted hours when combined with information on days worked and actual hours worked last week; this series reports the midpoint of the two bounds. Assisted hours are zero by construction for non-users. Data comes from the Real-Time Population Survey (RPS), a nationally representative online labor market survey of working-age adults aged 18-64 in the United States. When using this work, please cite: The Rapid Adoption of Generative AI, A. Bick, A. Blandin, and D. Deming, Management Science (2026) (https://pubsonline.informs.org/doi/10.1287/mnsc.2025.02523). See https://www.genaiadoptiontracker.com/ for more information. |