OpenAI's 25-city tour reveals a world hungry for AI but wary of its risks

After a month-long listening tour across 25 cities on six continents, OpenAI is sharing what it heard—and it’s a mixed bag of genuine excitement and deep-seated anxiety. The company, led by CEO Sam Altman, didn’t just drop in for photo ops. They sat with high school students in Nigeria using ChatGPT to dissect complex topics, civil servants in Singapore streamlining public services, and a French grocery chain using the tools to cut food waste. These aren’t hypothetical use cases; they’re live deployments that show the technology is already reshaping work and learning in tangible ways. The creativity, OpenAI notes, was genuinely inspiring.

The consistency of the concerns was just as striking as the enthusiasm. From city to city, the same fears bubbled up: misinformation, economic displacement, and the safety risks of increasingly powerful models. People aren’t just worried about some distant AGI nightmare. They’re worried about the models shipping today. OpenAI heard this loud and clear and saw an opportunity to clarify its data practices, reiterating that it doesn’t train on API customer data and that ChatGPT users can easily opt out. It’s a message they clearly felt needed repeating on a global stage.

Policymakers, unsurprisingly, were deeply engaged and remarkably aligned in their message. Dozens of heads of state and senior officials delivered a consistent demand: maximize the benefits for citizens, but put real guardrails in place—both for today’s tech and the more powerful systems on the horizon. They want ongoing dialogue and firm safety commitments from labs like OpenAI, and they’re open to exploring a global framework for governing future models. It’s not a regulatory free-for-all, but a clear signal that the era of self-regulation has an expiration date.

In response, OpenAI is placing new emphasis on making its products truly global, not just available in other languages. That means working on models that reflect local cultures and contexts, not just perform well on English-centric benchmarks. They’re also committing to refining concrete governance practices for highly capable foundation models and keeping pricing accessible for developers worldwide. The tour didn’t reveal a single, neat answer, but it gave OpenAI a clearer picture of the tightrope it has to walk—delivering powerful, useful tools while convincing a skeptical world it can handle the responsibility.