What people are really using ChatGPT for and why it’s not what you think

It was the most downloaded app in the world last month, one in 10 adults globally now use it, and people are increasingly saying they can’t live without it. But until now, no one really knew what people were doing with ChatGPT.

Most of the accounts were anecdotal: people using it to reply to emails, do homework or spiral into an AI-induced psychosis. Other use cases were fed to us by the company that created it, OpenAI. Chief executive Sam Altman said earlier this year that users were turning to it for legal help, medical questions and life advice, though no indication was given as to how niche such applications were.

This week, the first ever comprehensive study on how people actually use ChatGPT was published by OpenAI’s Economic Research team, who worked together with Harvard economist David Deming to analyse 1.5 million conversations between users and the AI chatbot.

The study revealed that more than three quarters of all conversations relate to practical guidance, seeking information and writing. In a blog post detailing the results, OpenAI claimed that the results showed that “access to AI should be treated as a basic right”.

The study found that of the 700 million people who now use ChatGPT every week, the number one request is for specific information. It demonstrates how generative artificial intelligence is increasingly usurping traditional search engines like Google, while also serving a much broader role as a personal assistant that can complete tasks and offer advice.

OpenAI said the findings “underscored ChatGPT’s dual role as both a productivity tool and a driver of value for consumers in daily life”.

The company added: “ChatGPT helps improve judgment and productivity, especially in knowledge-intensive jobs. And as people discover these and other benefits, usage deepens – with user cohorts increasing their activity over time through improved models and new use-case discovery.”

open image in gallery

The findings not only revealed how people were using ChatGPT, but the evolving way in which it is being integrated into people’s lives.

The headline revelation was that most consumers use it for non-work purposes, with only around 30 per cent turning to it in a professional capacity. But digging deeper into the data, certain patterns emerged about three distinct ways that people approach ChatGPT, categorised as “asking, doing, and expressing”.

Just over half of all messages were for “asking”, while “doing” accounted for just 35 per cent of requests – down more than 10 per cent from the year before. This comes despite OpenAI’s previous claims about how increasing work output was central to ChatGPT’s function. The firm continues to push this narrative, with the phrase “economic value” appearing five times in the 10-paragraph blog post.

Only around 14 per cent of interactions were for “expressing” activities, like self-reflection, however this has grown significantly over the last year.

The study also revealed stark differences in the way that men and women use ChatGPT.

By analysing users with typically masculine names and comparing their habits to users with typically feminine names, the researchers discovered that women are more likely to use the AI tool for practical guidance and writing assistance. By contrast, men more often use it for technical help and multimedia.

open image in gallery

Men are also now the minority of users, despite making up nearly two thirds of early adopters. Another demographic trend revealed in the 64-page study related to the age of users, with people under the age of 26 accounting for around half of all ChatGPT consumers. Though again this age gap is also levelling out over time.

“Usage gaps are closing as we increasingly democratise AI,” OpenAI’s blog post stated. “As of mid-2025, ChatGPT’s early gender gaps have narrowed dramatically, with adoption resembling the general adult population.”

The newly published data may offer a first-of-its-kind insight into generative AI, but the technology is still not fully formed, and we still do not know all the ways it might change the way we live, work and experience the world around us.

One of the most telling revelations about the way we treat ChatGPT was perhaps from one of the study’s footnotes. It referenced a separate study by researchers at Carnegie Mellon earlier this year that found that the average user in the US is so reliant on generative AI that they would have to be paid $98 to forgo it for a month.

This is a startling price for a technology that didn't even exist three years ago. But rather than express concern about over-dependence or even addiction to the AI chatbot, OpenAI described it as a $97 billion business opportunity. For a company that is still a long way from making a profit, this will likely be a key focus in how it develops the app going forward, and in turn how people will use it.