
What to know
- Check who qualifies as a user before comparing adoption figures.
- Ask about additional checking and correction work when evaluating a time-saving claim.
What the survey measures
The worker base was 20,275, including employees, self-employed people and gig workers. The report counts only monthly-or-more workplace users as AI users.
Hours saved were asked only after respondents selected a benefit. Nonusers, once-or-twice users, users reporting no benefit and missing answers count as zero. The survey did not ask about time lost to AI. These are self-reports, not measured productivity gains.
Sources: Northeastern University IIID
Our interpretation
When evaluating AI at work, include checking and correction time as well as drafting time.
Sources & method
June 21–August 11, 2026: 38,375 US adults; PureSpectrum online nonprobability sample, demographic quotas and population weighting. The landing page is dated November 18; the PDF says released November 17.
- AI and Time at Work Northeastern University IIID ·
- Report landing page Northeastern University IIID ·
Drafted and source-checked by Codex AI agents. Read our sourcing, illustration, and correction policy.
Explore the occupations
What could this mean for your work?
See task exposure and potential career moves for roles connected to this story.
