The World Is Talking to ChatGPT: What the Data Really Shows About AI in Daily Life
Introduction: The AI Conversation We’re Already Having
The numbers are hard to ignore. Every second of every day, tens of thousands of people are having conversations with ChatGPT. Over 2.5 billion messages daily, more than 29,000 every second, flow through the system. That makes ChatGPT not just the fastest-spreading technology of our lifetimes, but arguably one of the most widely used tools on the planet.
And yet, there’s still this strange disconnect. When I stand in front of a crowded room and ask who uses AI regularly, only half the hands go up. The reality is that even if many don’t see themselves as “AI users,” the world is already relying on it. AI is sitting in search boxes, editing suites, research workflows, classrooms, and living rooms.
So what are people really doing with ChatGPT? A new study by researchers at OpenAI, Harvard, and Duke gives us the clearest picture yet. The findings are fascinating — and the implications go far beyond productivity stats or corporate efficiency.
The Explosive Growth of ChatGPT
When ChatGPT launched in late 2022, few could have predicted just how quickly it would take off. Within five days it hit its first million users. Within a year, it had 100 million. By mid-2025, it was being used weekly by more than 700 million people, nearly 10% of the world’s adult population.
And it isn’t just a Western story. Growth has been particularly strong in low- and middle-income countries, where adoption rates are accelerating fastest. This suggests AI is not a luxury add-on for knowledge workers, but an increasingly global utility.
Here’s what stands out to me: we’ve seen rapid adoption curves before — smartphones, social media, streaming — but nothing matches this pace. And the genie is out of the bottle. Once hundreds of millions of people build habits around a tool, there’s no putting it back.
What People Are Really Doing With AI
OpenAI’s classification of conversations paints a revealing picture:
- Practical Guidance (28–29%): tutoring, how-to advice, creative ideas, health and fitness tips.
- Seeking Information (24%): straight factual queries, often substituting for Google.
- Writing (24%): emails, reports, translations, editing existing drafts.
Together, those three categories make up nearly 80% of all usage.
This alone is telling. The cliché is that AI is either for coders or for companionship, but the data disagrees. Programming is just 4.2% of ChatGPT messages, and social-emotional use is under 2%. That’s much smaller than reports on Anthropic’s Claude or broader “GenAI” surveys, where coding and companionship often rank higher.
The truth is far more practical: people use AI to get stuff done and to answer questions. It’s a tool for thinking, deciding, and producing.
Asking, Doing, Expressing: Three Modes of Use
Researchers also sorted usage into three simple categories of intent:
- Asking (51.6%): information-seeking, decision support, clarity.
- Doing (34.6%): delegating a task — writing, coding, designing.
- Expressing (13.8%): chatting, reflecting, or playing.
That split has changed over time. In 2024, Asking and Doing were almost even. By mid-2025, Asking had grown fastest. People are increasingly treating ChatGPT like a thinking partner, not just a task rabbit.
At work, though, Doing still dominates. About 56% of work-related queries are “Doing” — and most of those are writing. Think about that: in offices across the world, AI has quietly become the invisible co-writer of emails, reports, policies, and presentations.
The Shrinking Role of Work — And the Rise of Everyday Use
One of the most surprising findings is how non-work use is outpacing professional use. In June 2024, about half of ChatGPT messages were work-related. By June 2025, that share had dropped to just 27%.
It’s not that people stopped using AI at work. It’s that they’re using it even more outside of it. Recipes, workouts, tutoring, personal projects — everyday life is now an AI playground.
This matters. When people learn to use AI in their personal lives, they carry that fluency back into the workplace. It’s reminiscent of how personal smartphone adoption outpaced enterprise IT policies: individuals drove the change, and businesses scrambled to catch up.
Who Uses ChatGPT? The Demographic Shifts
The study also gives us a breakdown of who’s driving usage:
- Gender: Early users were overwhelmingly male (about 80%). By mid-2025, the split had not just balanced but tipped slightly female. Women tend to use ChatGPT more for writing and guidance, men for technical and multimedia tasks.
- Age: Nearly half of all usage comes from people under 26. Older users skew more toward work-related queries, but every age group is leaning more toward non-work over time.
- Education: Higher education correlates with more work use. Graduate-level users are the most likely to send “Asking” queries, suggesting they value AI most as a research assistant.
- Global distribution: Growth is strongest in low- and middle-income countries, widening the reach of generative AI beyond the traditional hubs.
This is an important reminder: AI is not just a Silicon Valley toy or a managerial assistant. It’s becoming a global commons, shaped by diverse voices and needs.
Individuals Are Outpacing Institutions
Here’s the striking part for me. People are running ahead of companies.
Employees are teaching themselves AI skills, experimenting in their own time, and applying them to side projects. They’re learning to draft websites, launch apps, automate spreadsheets, or streamline schoolwork — often without waiting for their employer’s blessing, or going it alone and starting up their own new venture without the “weight” or “digital debt” of the business they work for.
Meanwhile, businesses are split:
- Some are embracing it well, with training, authorised tools, and strategy.
- Others are embracing it badly, cutting staff too quickly and then having to reverse course (Klarna is a textbook example).
- Many are stuck: no policy, no training, no leadership.
- And some are ignoring it altogether. Those ones will fail.
The lesson for leaders is simple: your employees may already know more about AI than you do. If you don’t harness that, they’ll either leave or build something on their own.
A Personal Lens: Upskilling in Action
We’ve been working on this shift directly. At the end of this month, we’ll launch our online course, Practical AI for Everyday & Business Tasks. It’s designed as 20 hours of hands-on CPD — not abstract theory, but concrete ways to integrate AI into daily life and work.
As a test, I gave the course to my mother, who had never touched AI before. Within a few sessions, she’d gone from hesitant to empowered. She meal-planned with AI, rewrote her CV, drafted policies, and even generated illustrations for a children’s story she’d abandoned years ago.
Her words stuck with me: “I thought this wasn’t for me. Now I see it will change my life.”
If she can find that value, anyone can. And that’s the point: AI is no longer about early adopters. It’s about everyone.
Find out more about our brand new course and the work we have been doing in educating the South West here - https://tsrs.education/ai
The Strategic Implications
For businesses:
- Don’t underestimate personal adoption. Your employees are learning AI outside work, whether you sanction it or not.
- Focus on writing. It’s the most common work-related use case across all professions. Training here pays off immediately.
- Think decision support, not just automation. AI’s biggest value is helping people make better calls, not just spitting out documents.
For individuals:
- Experiment daily. The more you try AI for small things, the faster you’ll build fluency.
- Upskill intentionally. Treat it as a core career skill, like email or Excel.
- Share your learnings. Often the most innovative uses come from the grassroots, not the boardroom.
Closing Thought: The Next Shift
AI is already part of daily life. But here’s the kicker: we haven’t yet seen what happens when the billions of casual users — the students meal-planning, the parents tutoring, the freelancers tinkering — turn their skills fully onto business challenges.
That wave is coming. The question is whether leaders will ride it or be swamped by it.
Are you ready?
Source
How People Are Using ChatGPT - NBER Working Paper - View Paper