OpenAI Signals
A data insights product released by OpenAI on ChatGPT usage. It classifies and aggregates de-identified, aggregated platform logs at the task level by work activity, geography and time, and is used to analyze patterns and trends in AI use in work settings.
Timeline (last 90 days)
The OpenAI Signals dataset covers 165 work activities.
The OpenAI Signals dataset spans nearly 100 countries.
The core of OpenAI Signals is task-level classification and statistics on "what people do with ChatGPT," rather than financial or user growth metrics.
OpenAI Signals data typically comes from de-identified aggregated logs on the platform side, summarized by dimensions such as work activity, geography, and time.
The time dimension of OpenAI Signals provides an observation window for distinguishing novelty-seeking use from stable habits.
When 165 work activities are counted individually, it means AI penetration has moved beyond coarse scenarios like "writing copy" into more specific workflow stages.
OpenAI Signals reflects usage within the ChatGPT platform, cannot represent the entire AI tool market, and is not directly equivalent to quantitative evidence of productivity gains.
The classification framework of 165 work activities is set by the platform, which may amplify certain easily identifiable tasks (such as writing emails and debugging code) while underestimating composite work that is hard to categorize.
All Facts (8)
The OpenAI Signals dataset covers 165 work activities.
50%UnverifiedThe OpenAI Signals dataset spans nearly 100 countries.
50%UnverifiedThe core of OpenAI Signals is task-level classification and statistics on "what people do with ChatGPT," rather than financial or user growth metrics.
50%UnverifiedOpenAI Signals data typically comes from de-identified aggregated logs on the platform side, summarized by dimensions such as work activity, geography, and time.
50%UnverifiedThe time dimension of OpenAI Signals provides an observation window for distinguishing novelty-seeking use from stable habits.
50%UnverifiedWhen 165 work activities are counted individually, it means AI penetration has moved beyond coarse scenarios like "writing copy" into more specific workflow stages.
50%UnverifiedOpenAI Signals reflects usage within the ChatGPT platform, cannot represent the entire AI tool market, and is not directly equivalent to quantitative evidence of productivity gains.
50%UnverifiedThe classification framework of 165 work activities is set by the platform, which may amplify certain easily identifiable tasks (such as writing emails and debugging code) while underestimating composite work that is hard to categorize.
50%