The world thinks China is winning the AI race. It just doesn’t trust the winner.

Ask people around the world who is winning the AI race, and a surprising number now answer: China.

In 11 of the 15 countries polled for a new report by the London consultancy Public First, a clear majority said they believe China is now ahead of the United States on AI capability and innovation.

The survey covered more than 18,000 people, and the list of countries that ranked China first included some of America’s closest allies: Canada, Britain and France among them.

A verdict on capability

This is a shift in perception, not a measurement of the technology itself. But perception matters, and it tracks a real narrowing. On the hard benchmarks, the gap between the top US and Chinese models has shrunk to a few percentage points, as Stanford’s AI Index found this year.

The public seems to have absorbed the headlines about cheap, capable Chinese models, and concluded the race is closer than Washington would like.

The poll spanned both advanced economies, including Canada, France, Germany, Japan and the UK, and emerging ones such as Brazil, India, Indonesia and Nigeria. The view that China leads held across both groups.

Capability is not trust

Here is the twist. Being seen as the most capable is not the same as being the most trusted. When Public First asked whose models people actually trusted, the answer was messier.

People trusted models from their own country first. After that, Japan came out as the most trusted source overall, with the US just behind. China, despite its lead on perceived capability, did not top the trust rankings. Instead it split opinion, drawing both strong trust and strong distrust.

The US was not immune either: 19 per cent of non-Americans said they did not trust an American model.

That gap between admiration and trust is the whole story. People can believe China builds the best AI and still not want to hand it their data.

A Western backlash is brewing

The trust problem sits on top of a souring mood in the West. In the US, the balance of positive over negative feeling about AI has collapsed, from a strongly positive score two years ago to barely positive now. In the UK, net positivity among 18-to-24-year-olds fell from 62 per cent in 2023 to 46 per cent in 2026. Asian and emerging markets stayed far more optimistic, echoing the growing gap between AI insiders and everyone else.

Why it matters

The findings point to a world that wants AI on its own terms. More than three-quarters of respondents said it was important to keep their data inside their own country, a strong vote for digital sovereignty as governments fight over whose models the rest of the world gets to use.

But the same people are pragmatic. Of those who said they distrusted US models, most were still using ChatGPT, Claude or Gemini anyway. People may believe China is winning, and may want their data kept at home, but for now they still reach for whatever works.