URGENT UPDATE: OpenAI has just released groundbreaking findings on how users are engaging with ChatGPT, revealing a significant shift in personal versus work-related usage. In a study conducted with the National Bureau of Economic Research (NBER), the data shows that approximately 80% of ChatGPT interactions fall into three main categories: “Practical Guidance,” “Seeking Information,” and “Writing.”
The research indicates a remarkable increase in personal use, with non-work-related interactions skyrocketing from 53% to 73% over the past year. Meanwhile, work usage has plummeted from 47% in June 2024 to just 27% in June 2025. This trend highlights a broader struggle within companies to effectively leverage AI technologies for substantial returns.
The study, authored by OpenAI’s Economic Research team alongside Harvard economist David Deming, categorizes the most common uses of ChatGPT. “Practical guidance” includes activities such as tutoring and creative ideation, while “seeking information” serves as a modern alternative to traditional search engines. Notably, “Writing” emerged as the leading work-related function, constituting 40% of job-related messages in June 2025, starkly contrasting with a mere 4.2% for programming tasks.
Despite the growing prevalence of personal use, OpenAI discovered that only a “fairly small” percentage of users engage with ChatGPT for social-emotional support. Approximately 2% of messages were from users seeking companionship or discussing personal issues, raising questions about the depth of engagement and the impact of AI in mental health support. Contradicting these findings, other research suggests that up to one in three teenagers utilize AI chatbots for social interaction, highlighting a potential disconnect in user experiences.
The demographic shift is also notable. OpenAI’s research indicates that 46% of messages are generated by users aged 18-25, a group that favors personal interactions with the chatbot. Furthermore, there is a closing gender gap in ChatGPT usage; the proportion of users with “masculine first names” has dropped from 80% in 2022 to 48%, while users with “typically feminine names” are now reaching parity.
Experts urge caution regarding the reliability of these findings, as OpenAI employed AI to categorize user messages. However, the implications of this data are significant, indicating a shift in how individuals, especially younger users, perceive and utilize AI technology.
As ChatGPT continues to evolve, the growing trend toward personal engagement raises critical questions about the future of AI interactions. What does this mean for users seeking emotional support? Will companies adapt to this changing landscape?
Stay tuned for further developments as OpenAI and researchers continue to explore the evolving dynamics of AI usage globally.
