The Information Machine

How AI is expanding what people do at work

OpenAI Blog · 2026-07-27

OpenAI Economic Research analyzes over 800,000 work-related ChatGPT messages to find that 43.5% of occupation-specific AI use involves tasks from a different field, a pattern called 'task crossover' suggesting AI is reorganizing who does what at work before those changes appear in job titles or labor statistics.

Open original ↗

Extraction

Topics: ai-labor-economicsfuture-of-worktask-crossoverai-productivityoccupational-change

Claims

  • In an analysis of over 800,000 work-related ChatGPT messages, 43.5% of occupation-specific messages involved tasks normally associated with a different occupation.
  • Customer experience workers (77%), designers (75%), and HR workers (69%) showed the highest rates of using AI for tasks outside their own occupational domain.
  • Marketing tasks spread most broadly across other occupations, appearing in 8.9% of messages from workers in non-marketing fields — the highest outward share in the sample.
  • Workers in smaller organizations (2–5 seats) showed higher outside-occupation task crossover at 18.9% versus 16.3% for those in organizations with over 100 seats, suggesting AI serves as a generalist tool where specialists are scarce.
  • AI usage patterns may serve as a leading indicator of occupational change that conventional labor-market statistics will capture only later.

Key quotes

AI changes not just how work gets done, but who does what.
AI may be especially useful as a generalist tool where specialist resources are scarce.
Usage patterns may provide an early signal of occupational change that conventional labor-market statistics will capture only later.