OpenAI has partnered with researchers from Duke University and Harvard University and published a new study titled “How People Use ChatGPT.” It gives insights into how the chatbot has grown and what people do with it.
The paper tracks ChatGPT’s rise from its November 2022 launch to July 2025, when the service reached 700 million weekly active users, which is about 10% of the world’s adult population, sending 2.5 billion messages a day, or roughly 29,000 messages every second. According to the authors of the study, no consumer technology has spread this quickly on a global scale.
The researchers found that non-work usage has surged on ChatGPT, growing from 53% of messages in mid-2024 to 73% by mid-2025. While generative AI is often discussed as a workplace productivity tool, the report suggests its impact on personal life and home activities is at least as significant.
However, there are three categories that contribute to about 78% of all conversations. It includes Practical Guidance, Seeking Information, and Writing. Tasks like writing dominate professional use, with two-thirds of those requests focused on editing or improving text rather than creating something from scratch. Apart from that, computer programming accounted for just 4.2% of queries, and relationship or companionship chats made up only 1.9%, which is the opposite of common assumptions about how the tool is used.
Furthermore, the gender balance of users has shifted dramatically. In the first months after launch, about 80% of active users had typically masculine first names. However, by mid-2025, women slightly outnumbered men among weekly users. The study also found that nearly half of all messages come from users under age 26, and growth has been fastest in low- and middle-income countries. We have already seen that India has become the second-largest market for ChatGPT, and as a result, OpenAI is investing heavily in the country.
The authors also emphasised that ChatGPT is being used across industries for similar tasks like gathering information, solving problems, and making decisions, regardless of profession. They argue that generative AI’s ability to provide decision support and produce digital outputs like writing gives it economic value far beyond traditional search engines.
