2 min read

Americans Think China Is Winning the AI Race

Americans Think China Is Winning the AI Race

A new Pew Research Center survey of 3,488 U.S. adults, conducted June 22-28, finds that by a three-to-one margin, Americans say China is more advanced than the United States in developing AI. Only 12% say the U.S. leads; 36% say China does; 18% call it roughly even; a third aren't sure.

Key Points

  • 36% of Americans say China is more advanced in AI development, compared to 12% who say the U.S. leads, with 18% calling it about equal and 33% unsure
  • 43% say it's extremely or very important that the U.S. lead the world in AI development; 22% say it's not too or not at all important
  • 54% of Republicans say U.S. AI leadership is important, compared to 34% of Democrats
  • Men are more than twice as likely as women to say the U.S. leads China on AI (18% vs. 7%)
  • 51% of Americans say AI will widen the gap between rich and poor countries, versus 16% who think it won't make much difference and 7% who expect it to narrow the gap

Perception Doesn't Match the Usual Industry Narrative

American AI labs regularly frame themselves as ahead, and coverage of frontier model launches tends to reinforce that framing. The public isn't buying it uniformly. Nearly three times as many Americans think China has the edge as think the U.S. does, and a third simply don't know. That gap between industry messaging and public perception is worth sitting with, especially for any brand whose positioning leans on "American AI leadership" as a selling point.

The Divides Run Along Predictable Lines, With One Exception

Political affiliation splits the "importance" question sharply: 54% of Republicans call U.S. AI leadership important versus 34% of Democrats, consistent with the Trump administration's public framing of AI as a dominance race. But the "who's actually ahead" question doesn't split the same way. Majorities of both Republicans and Democrats say China is more advanced, even as they disagree on how much U.S. leadership should matter. Believing the U.S. should lead and believing it currently does are two different questions, and Americans across the political spectrum are answering them differently.

People Who Use AI Chatbots See the Race Differently

Americans who use AI chatbots are more likely than non-users to say the U.S. is ahead (16% vs. 9%), but the two groups land on the same "China is more advanced" number (36% each). Non-users are simply more uncertain overall. That's a useful signal for anyone messaging to a general audience: familiarity with AI tools doesn't necessarily produce more optimism about American competitiveness, it mostly produces more confidence in an opinion either way.

How Brands Talk About AI Right Now

If a majority of your audience already believes China is ahead on AI development, and 51% expect AI to widen global inequality, messaging that leans on unqualified American AI exceptionalism is working against the current, not with it. That's a positioning question worth resolving deliberately rather than defaulting into, and it's exactly the kind of audience-perception work that belongs in a growth strategy conversation before a campaign goes out, not after it underperforms.

For teams building AI-related messaging into 2026 campaigns, understanding where public sentiment actually sits, not where industry press assumes it sits, is worth the extra diligence. That's the kind of audience research our AI marketing services team folds into campaign planning.

Alibaba Says Its New Model Is Second Best In The World

Alibaba Says Its New Model Is Second Best In The World

Alibaba previewed Qwen 3.8-Max at the World Artificial Intelligence Conference in Shanghai and made a specific claim: the 2.4-trillion-parameter...

Read More
the Government Controls Access to OpenAi's New Models

the Government Controls Access to OpenAi's New Models

Two weeks ago, Anthropic's most capable models went dark under a government export control directive. On June 26, OpenAI launched three new models...

Read More
The Best AI Models of 2025: When

The Best AI Models of 2025: When "Best" Stopped Meaning Anything

Decrypt's year-end AI model roundup reveals the defining insight of 2025: nobody's using a single "best" model anymore. They're assembling...

Read More