ChatGPT at Work: What the New Data from OpenAI Reveals

OpenAI has just released, through its OpenAI Signals platform, its first detailed country-by-country statistics on ChatGPT usage. The clearest signal: at work, people are no longer just asking questions — they’re more than twice as likely to have ChatGPT produce something, from writing and code to analysis, than they are outside of work. Three other trends round out the picture: adoption is spreading into regions that had been lagging (Latin America, Africa, Oceania), multimedia use is growing faster than any other use case, and people over 35 are closing the adoption gap, with France posting one of the sharpest increases in Europe.

From a tool you ask to a tool that produces

This is the most significant shift in the new data. In a professional context, ChatGPT users are more than twice as likely to “do” — meaning they ask for a deliverable or a completed task (writing, coding, data analysis) — than to simply look something up. Outside of work, the pattern flips: exploratory use, asking to understand, is still the dominant mode.

This isn’t a minor detail for anyone running teams or digital tools: it reflects a change in posture. Generative AI is no longer seen mainly as an answer engine — it’s becoming an execution layer built into the workflow itself, a silent collaborator handed a task rather than a question.

Global adoption is rebalancing

The second finding concerns the geography of adoption. Countries that started out with lower per-capita usage are gradually catching up to early adopters (North America, Western Europe, parts of Asia). In the second quarter of 2026, countries in Latin America, Oceania, and Africa climbed fastest in the global per-capita rankings — with Peru, Uruguay, and Costa Rica posting the biggest jumps.

North America and Europe keep growing too, but the gap is narrowing: generative AI is steadily moving beyond its original circle of early adopters to become a more universal tool.

Multimedia: the fastest-growing use case

Another trend worth watching: the share of conversations involving multimedia generation, analysis, or retrieval (images in particular) is the one growing fastest, now accounting for roughly 7.8% of messages worldwide. That’s still behind the long-standing top use cases — practical guidance, writing, information-seeking — but it has clearly accelerated since a new generation of image tools launched in spring 2026.

The effect is especially pronounced in Latin America: in Brazil and Colombia, more than one message in ten is now classified as multimedia.

Users over 35 are closing the generational gap

One last signal, and an important one for anyone tracking digital-tool adoption inside companies: the share of messages sent by users over 35 rose in nearly every country studied, up an average of 5 percentage points year over year. This reading is based on users who voluntarily reported their age on the platform.

In Europe, France and Czechia stand out in particular: the share of messages from users 35 and older rose by more than 10 percentage points over the past year, a faster pace than the continental average — a sign that generative AI has moved well beyond the tech-savvy, younger crowd that drove its early adoption.

What this means for e-commerce and digital teams

Three practical implications stand out from this global snapshot:

• Treat these tools as executors, not just answer engines. If professional use is shifting toward “doing,” the use cases worth prioritizing are the ones that produce something directly usable: draft product descriptions, first-pass reporting, automation scripts.

• Stop treating AI as a junior or “tech-friendly” tool only. The growth among users 35 and older, especially in France, suggests training and support now need to reach the whole team, not just early adopters.

• Plan for the rise of multimedia. Image generation and analysis are likely to show up more in e-commerce workflows — product visuals, moderation, competitive content analysis — and will deserve the same level of governance that text already gets.

Conclusion

This data confirms a trajectory that was already taking shape: generative AI is moving away from being a tool you consult and toward being a tool you build into the production flow — across a growing number of countries, generations, and formats. For organizations, the question is no longer whether adoption will become widespread, but how to structure the usage, training, and governance that need to come with it, starting now.

Source: OpenAI, From asking to doing: How the world is putting ChatGPT to work, OpenAI Signals, August 6, 2026. Read the original article - This article was written with the assistance of an AI model (Claude, Anthropic), based on an analysis of OpenAI’s original publication.