In 2025 ChatGPT it has become a faithful mirror of the cognitive, working and relational habits of hundreds of millions of people. According to data analyzed by the National Bureau of Economic Research (NBER) on a sample of 1.1 million conversations between May 2024 and June 2025, over 800 million users around the world they interact weekly with generative artificial intelligence chatbots, and the way they do it tells much more than a simple technological evolution.
In particular, the use of ChatGPT is mainly concentrated in three large areas: writing, practical guidance and information seeking. Together, these categories cover over 75% of interactionsconfirming that generative AI is now an everyday tool to work better, understand more and communicate more effectively.
Write better (and faster): The first big use of ChatGPT
The most relevant category is writingwhich alone represents 28% of requests. It’s not just about literary creativity (which weighs just 1%) but above all about functional and professional use.
The main item is “edit or critique provided text” (11%), i.e. “modify or evaluate the text provided”. That is, users ask ChatGPT to correct, improve, simplify or make more effective a text already written. This is followed by personal writing (emails, messages, written requests) and communications (8%), translations (5%) and synthesis or generation of arguments (4%).
ChatGPT therefore does not replace human writing, but optimizes it. It is a cognitive accelerator that reduces the time needed to structure a message, adapt the tone to a professional context, reformulate content for social media or for work. And it is precisely this function of “silent co-author” that explains how and why AI has now firmly entered the workflows of journalists, consultants, students and companies.
Practical guidance and tutoring: AI as a daily consultant
And other 28% of conversations falls within the macro-area of practical guidance. Here ChatGPT is used as a real one consulente on demand. Tutoring and teaching (10%) are the main item, followed by practical advice on how to do something (9%) and by requests about health, fitness, beauty and self-care (6%). Requests for creative ideas close (4%).
This use is consistent with the vision described by great AI leaders. Jensen Huang (Nvidia) and Masayoshi Son (SoftBank) insist that AI will make human work more productive and meaningful by freeing up time from repetitive tasks. However, in practice, ChatGPT is already used as personalized tutorespecially in the educational sector, where the first cracks emerge. According to the College Board, 84% of US high school students use generative AI for i schoolwork. Not always to study better. Often for delegate critical thinking.
Searching for information: less Google, more dialogue
The search for information accounts for 21% of requests. The “specific information” item alone is worth 18%. People use ChatGPT to get timely answers, contextualized explanations and immediate summaries. Only 2% concern purchasable products, a sign that, at least for now, AI is perceived more as a cognitive tool than a commercial channel.
This behavior suggests a structural change, that is, we are no longer just looking for information, but for intelligent interlocution. ChatGPT does not return links, but reasoning. And this is precisely what makes it more powerful than Google search rankings.
Technology, multimedia and self-expression: growing niches
Then there are the technical requests (8%), especially on programming (4%) and mathematical calculations (3%), and multimedia ones (6%), which concern the image creationi (4%). But there is also room for the so-called “self-expression” (4%), which includes personal reflections, relationships and even simple informal conversations.
This is where one of the most controversial topics comes in, namely the optimization of engagement. Chatbots, like social networks, are designed to detain the user. Usage data analysis tells how AI systems, including ChatGPT, are not only answering questions, but also modeling human behaviors. In fact, they are changing the way we write, study and make decisions, expanding both human capabilities and frailties.
On the one hand, the promise is enormous: productivity, efficiency, accelerated scientific research, personalized medicine, intelligent supply chains. On the other hand, there are growing concerns about the concentration of power, the impact on employment, the energy consumption of data centers and the cognitive dependence.
