From asking to doing: How the world is putting ChatGPT to work
Quick Answer
OpenAI's data reveals a significant shift in ChatGPT usage from inquiry to action, particularly in professional settings where task completion is prioritized.
Quick Take
Global adoption is increasing, especially in Latin America and among users over 35, with multimedia applications growing rapidly, now comprising 7.8% of messages.
Key Points
- Users are over twice as likely to use ChatGPT for tasks at work than outside.
- Adoption rates are rising in Latin America, Africa, and Oceania, narrowing the gap with early adopters.
- Multimedia use is the fastest-growing category, accounting for 7.8% of messages globally.
- Users over 35 are sending 5% more messages compared to last year, with significant growth in France and Czechia.
- The Signals dataset provides insights into how over 1 billion users are leveraging ChatGPT.
DeepSignal Analysis
What happened
OpenAI's recent data indicates a notable transition in ChatGPT usage from inquiry to task completion, particularly in professional environments. The adoption of ChatGPT is expanding globally, especially in Latin America and among users aged 35 and older, with multimedia applications also on the rise, now representing 7.8% of messages.
Key evidence
- Users are more than twice as likely to utilize ChatGPT for task completion at work compared to outside of work, where exploratory inquiries dominate.
- Countries in Latin America, Africa, and Oceania are experiencing rapid growth in ChatGPT adoption, with notable increases in Peru, Uruguay, and Costa Rica.
- The share of messages categorized as multimedia has risen to 7.8%, with countries like Brazil and Colombia showing over 10% of messages focused on multimedia.
Why it matters
This shift in usage patterns suggests that ChatGPT is becoming a more integral tool for productivity, particularly in professional settings. The increasing adoption among older users and in diverse regions indicates a broadening appeal and utility of AI technologies. Understanding these trends can help policymakers and businesses adapt to the evolving landscape of AI integration.
📖 Reader Mode
~5 min readAI is moving from something people ask to something they put to work.
For the first time, OpenAI is publishing country-by-country data showing how people around the world use ChatGPT. The findings reveal a technology that is moving beyond answers and into action, spreading beyond early adopters, and becoming useful to more people in more ways:
- From asking to doing: At work, people are more than twice as likely to use ChatGPT to complete a task or create something, from writing and coding to analysis, than they are outside work.
- AI adoption is becoming more global: Countries across Latin America, Africa, and Oceania are catching up to early adopters, narrowing the global adoption gap.
- AI is moving beyond text: Multimedia is the fastest-growing use case globally, accounting for 7.8% of messages and more than one in ten in countries including Brazil and Colombia.
- AI is reaching beyond early adopters: Usage among people over 35 is rising in nearly every country. In France and Czechia, their share of messages increased by more than 10 percentage points in the past year.
Taken together, the data points to a clear shift: AI is no longer just helping people find answers. It is helping more people, in more places, get things done.
This new data is available to view or download on OpenAI Signals, our hub for data, research, and analysis, from the OpenAI Economic Research Team. These country-level statistics will help people, including policymakers and researchers around the world, better understand how more than 1 billion people are putting ChatGPT to work.
The Signals dataset specifically reflects messages sent within ChatGPT Free, Go, Plus, and Pro accounts—the group of accounts generally managed by individuals rather than organizations.
At work, “doing” dominates
The data reflect a trend that has appeared in recent headlines, especially since the rollout of ChatGPT Work: people are increasingly using the technology, particularly for professional pursuits. In these settings, users are more frequently “doing”—using ChatGPT to produce an output or perform a task.
This type of work includes applications such as editing, coding, or conducting an analysis. Our data shows that, across the globe, people are more than twice as likely to use ChatGPT this way at work than outside of work. Outside of work, use is more exploratory: “asking,” or seeking information and clarification, remains the largest category.
The adoption gap is closing
Continuing the steadily climbing trend worldwide documented in June, the countries and regions that began their AI journeys with lower ChatGPT per-capita adoption rates are gradually catching up to the established early adopters.
For Q2 2026, our team updated our usage rankings across countries (see previous quarterly rankings here). The chart below shows how the per-capita leaderboard changed since the first quarter of 2026.
Notably, in the second quarter, usage in parts of Latin America, Oceania, and Africa increased faster than usage in other parts of the world, with Peru, Uruguay, and Costa Rica rising the most among countries in global rankings. Adoption continues to rise in North America and Europe, but parts of the Southern Hemisphere are catching up in per-capita adoption rates.
Change in country messages-per-capita rank, Q1 2026 to Q2 2026 (positive = improved)
Note: For 144 countries, blue indicates an improvement in rank, brown a decline, and gray no change. Hatching identifies excluded or ineligible countries.
As new capabilities emerge, specific use cases command people’s attention and drive adoption. Take multimedia generation, analysis, and retrieval, for example. Since the release of ChatGPT Images 2.0 in April 2026, the share of messages across the world focused on multimedia use increased to 7.8%. While it still lags behind the leading use cases of practical guidance, writing, and information-seeking, multimedia use has been on a consistent upswing year-to-date.
Drilling deeper into the category of multimedia messages, the data shows distinctive differences among countries and regions. The graphic below demonstrates the uptick in Latin America where, in countries like Brazil and Colombia, more than one in ten messages to ChatGPT is categorized as multimedia.
Note: Share of classified individual ChatGPT messages categorized as multimedia, by country, in Q2 2026. The refreshed release includes 126 countries with an eligible Q2 multimedia estimate; striped countries are outside the release or have no eligible data.
ChatGPT adoption grew worldwide among people over 35
The quarter-over-quarter changes in per-capita adoption are not limited to use cases and regions alone. The data also demonstrate differences in adoption by age cohort.
The trend line over the last year shows that the percentage of messages sent by people over 35 increased in almost every country. Year over year, these users now account for a 5% higher share of messages than they did 12 months earlier.
This is an analysis of only users who self-reported their age on the ChatGPT platform.
Looking at the data country-by-country, adoption among people aged 35 and older is growing particularly quickly in several countries. The chart below, for example, shows how the share of ChatGPT messages sent by users 35 and above shifted significantly in several European countries. In France and Czechia, the share of messages sent by users 35 or older increased by more than 10 percentage points in the last year.
Almost three-quarters of European countries saw larger-than-average increases in their share of messages sent by these users, while the opposite pattern emerged in some Southeast Asian countries, such as Singapore. Overall, the share of messages sent by these users increased in six of the eight countries in the region, albeit by only a small margin.
Three-month trailing share of age-classified messages from users aged 35+, relative to each country’s Q2 2025 average (percentage points)
Note: Three-month trailing share of self-reported, age-classified individual messages from users aged 35 and older; changes are relative to each country’s own Q2 2025 average. Gray lines show all 111 countries with complete estimates; dashed gray shows the equal-country median. June 2026 endpoints compare the latest quarter with Q2 2025. Source: OpenAI Signals refreshed public-release CSVs.
More about OpenAI Signals
OpenAI Signals is the OpenAI Economic Research team’s platform for sharing our data and research with the world. In addition to our research, we publish regular updates on how individuals and businesses use ChatGPT. We encourage you to download the data here and explore our methodology here(opens in a new window).
— Originally published at openai.com
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from OpenAI Blog
See more →Scientific computing in the age of agentic AI
AI agents are transforming scientific computing by streamlining software development, enabling researchers to focus on discovery. Projects using Codex and Claude Code report accelerated development and improved maintenance, though challenges in validating AI outputs remain. Long-term stewardship of research software is crucial to ensure reliability and reproducibility.