How People Use ChatGPT

This study, conducted by OpenAI and Harvard economist David Deming, is the largest-ever report on ChatGPT usage patterns, analyzing 1.5 million conversations. The findings show that demographic gaps among users are rapidly narrowing due to the democratization of AI technology.

The gender gap in particular showed a notable decrease. In January 2024, users with feminine-coded names made up 37%, but by July 2025 this rose to 52%, becoming similar to the general population composition. Additionally, growth rates in low-income countries were more than 4 times higher than in high-income countries, solidifying ChatGPT's position as a global tool.

Usage patterns are broadly classified into three types:

• Asking (49%): Acting as an advisor. The stage of seeking information or requesting guidance.
• Doing (40%): Performing actual tasks such as drafting text, planning, and programming.
• Expressing (11%): Personal reflection, exploration, and play.

In terms of economic value, about 30% of ChatGPT usage was work-related, while 70% was related to everyday life. ChatGPT is generating real economic value that is difficult to capture with traditional GDP measurement methods, particularly by supporting decision-making and boosting productivity in knowledge-intensive jobs.