OpenAI has released aggregated data on work-related ChatGPT use across nearly 100 countries. Its October 26 research announcement maps usage to 165 Intermediate Work Activities, giving researchers a common vocabulary for comparing the work people bring to ChatGPT.
A quarterly view of work messages
The data spans the first quarter of 2025 through the third quarter of 2026. It draws on O*NET categories that group occupation-specific tasks into broader activities, such as preparing instructional material, programming and analysing business information.
OpenAI says researchers can compare patterns across countries, examine changes over time and connect the release with other economic indicators. The common categories make those comparisons possible, but they do not themselves show why usage differs.
What changed in the reported sample
Across 74 countries observed in every quarter, the ten most common activities accounted for an average 51.1% of classified work messages in the third quarter of 2026. The comparable figure was 56.0% in the first quarter of 2025.
That change describes concentration within the measured messages. It does not establish a change in national productivity or employment. OpenAI also notes that roughly 43% of the employment-weighted task mix for mapped US occupations sits outside the individually reported categories, including physical activities that may appear less often in ChatGPT conversations.
Coverage and privacy limits
The release covers individual ChatGPT plans and excludes enterprise messages and Codex use. Researchers do not read individual conversations. OpenAI says automated systems classify messages after identifying information is removed, then publish aggregates with small groups withheld and statistical noise added.
Those boundaries should travel with any chart made from the data. A comparison of consumer work messages can support questions about use patterns, while a claim about an entire workforce needs additional evidence and a method for connecting the two.