OpenAI Research: Over 40% of Occupation-Specific ChatGPT Use Crosses Job Boundaries
On July 27, 2026, OpenAI published a research report analyzing over 800,000 work-related ChatGPT messages from U.S. users, finding that 43.5% of occupation-specific tasks involved work outside a person’s own job function. The findings quantitatively confirm what OpenAI calls “task crossover” — professionals routinely using AI to do work traditionally associated with other roles.
Details
- Scope: Analyzed 800,000+ work-related ChatGPT messages from U.S. users, distinguishing generic tasks (writing, scheduling) from occupation-specific ones, then checking whether users were performing tasks outside their own occupation
- Overall crossover rate: 16.8% of all work-related messages, and 43.5% of occupation-specific messages, involved tasks outside the user’s own field. Excluding generic tasks, nearly half of occupation-specific usage crosses job boundaries
- Crossover by occupation: Customer experience workers topped the list at 77%, followed by designers (75%), HR professionals (69%), legal professionals (56%), and marketers (53%)
- Which tasks travel farthest: Marketing- and engineering-related tasks showed up most broadly across other occupations. Financial calculations and technology troubleshooting ranked among the top “outside” tasks across all seven occupation groups studied
- Company size effect: Smaller companies (2-5 employees) showed an 18.9% outside-task usage rate, versus 16.3% at larger organizations (100+ employees) — suggesting AI serves as a generalist tool where specialist staff aren’t available
How to try it
- This is a research report published on OpenAI’s site, not a new product or feature
- Teams can use the framework to review their own usage logs and identify which tasks are being requested across role boundaries, informing workforce planning and training
- The full report is available on OpenAI’s official site