OpenAI just released a research study of 1.5 million ChatGPT conversations and it provides insights into how people use the world’s most popular chatbot.
There’s a toss-up between “Writing” and “Practical Guidance” (about 28% each), followed closely by “Seeking Information” (21%). None of these figures are surprising, and to me, the data points to a future where:
🔷 The baseline level of writing and self-learning will go up and companies will demand higher standards of output from employees.
🔷 Traditional search will take a backseat to AI search.
Some might be surprised to see that “Relationships and Personal Reflection” usage is low at about 2%, but I suspect that people who use AI chatbots are turning to other dedicated apps.
“Create An Image” is pretty low at about 4%, despite all the Studio Ghibli memes. It’s probably because ChatGPT takes several minutes to generate an image, and it has been surpassed in speed and quality by Google Gemini.
“Data Analysis” is very low at under 1%. While I teach my workshop participants never to trust ChatGPT when doing math, I say it’s still good for doing a quick first pass on datasets, especially when you lack time or expertise on the data.
Other findings:
🟦 Gender gap closed: Majority of weekly users now have typically female first names.
🟦 Young users dominate: Nearly half of adult messages come from users under 26.
🟦 Rapid growth in low- and middle-income countries in the past year.
🟦 Education/occupation effect: Highly educated professionals more likely to use ChatGPT for work and for Asking (advisory) use cases.