Public opinion has shifted. In a striking new survey, respondents in 11 out of 15 countries now rank Chinese artificial intelligence models ahead of those from the United States on capability and innovation. Yet when it comes to trust, China falls short. Japan sits at the top. The U.S. follows close behind. China lands in 10th place with a net trust score of negative eight.
The findings come from a poll of more than 18,000 people conducted by London-based consultancy Public First. Coverage in The Next Web and South China Morning Post highlights the gap. More than 90 percent of those surveyed believe AI will transform the world. Fears linger anyway.
Capability perception tells one story. In Canada, Britain and France — key U.S. allies — at least 40 percent said outright that China leads. Only 23 percent in Germany still view the U.S. as ahead. Even inside the United States, 24 percent acknowledged China’s edge while 51 percent disagreed. The poll did not survey people in China itself.
Trust tells another. Respondents favored models from their own country first. Japan earned the highest overall confidence. The U.S. scored a net positive 16, good for second place. China’s negative eight reflected divided views. Strong pockets of trust existed alongside equally strong distrust. Nineteen percent of non-Americans said they simply would not trust an American model either.
But. The numbers reveal pragmatism. Many who expressed distrust in U.S. systems still used ChatGPT, Claude or Gemini daily. People reach for tools that work. They hesitate to hand over data to foreign powers.
More than three-quarters of participants called it important to keep their data inside their own borders. Digital sovereignty emerges as a priority. Governments wrestle over export controls and which models the world will use. The Public First report, detailed in recent coverage, captures this tension.
Domestic attitudes inside China paint a different picture. An Edelman poll from late 2025 found 87 percent of Chinese respondents trusted AI. That compared with 67 percent in Brazil, 39 percent in Germany, 36 percent in the UK and just 32 percent in the United States. Edelman noted hands-on experience accelerates acceptance. The more people use the technology, the faster trust grows.
Chinese respondents also expressed high optimism. Majorities expected generative tools to improve mental health, work and societal challenges like climate change. Western audiences showed far more caution. The contrast underscores two worlds of AI adoption.
Global surveys add layers. Pew Research in 2025 found median trust in China to regulate AI effectively stood at only 27 percent across 25 countries. The EU scored 53 percent. The U.S. landed at 37 percent. Americans themselves showed little faith in Beijing’s oversight. Just 13 percent trusted China on regulation. Seventy-six percent did not.
Stanford’s 2026 AI Index Report echoes the capability convergence. The performance gap between top U.S. and Chinese models has narrowed to a few percentage points. Public perception appears to have caught up with technical reality. Headlines about affordable, high-performing Chinese systems have registered worldwide.
Sentiment in the West has cooled. In the United States, net positive feelings toward AI dropped from strongly favorable two years ago to barely positive today. Among British respondents aged 18 to 24, that figure fell from 62 percent in 2023 to 46 percent in 2026. Asian and emerging markets remain upbeat. The divide between experts and the broader public grows.
Experience matters. Edelman’s flash poll showed large jumps in trust among those who reported AI helped them grasp complex ideas or boosted their work. In China and Brazil, over two-thirds felt confident in positive impacts. In Germany, the UK and U.S., trust hovered near or below 40 percent. “Trust will decide not only who adopts AI, but who benefits most from it,” an Edelman executive stated in the report.
Geopolitical rivalry colors the picture. Parallel surveys published in Global Policy journal in 2025 compared views in China and the U.S. Chinese participants expressed higher overall trust in various stakeholders, though both groups favored their own country. U.S. respondents proved more skeptical across the board.
Recent coverage reinforces the pattern. A KPMG study from May 2025 found Chinese respondents far more open to AI than the global average. Optimism reached 91 percent against worries at 44 percent. Emerging economies generally showed higher acceptance.
China Media Group released its own massive survey in June 2026 based on 100,000 responses. It pointed to widespread usage and confidence among the Chinese public on AI’s effects on work, education and skills. The domestic embrace stands in sharp contrast to international hesitation.
So what explains the trust deficit? History plays a role. Concerns over data practices, government oversight and potential misuse linger in many minds. Models developed in China split opinion precisely because capability does not erase those worries. Japan benefits from a cleaner image on technology and privacy. The U.S. retains an edge from established brands and alliances even as its lead narrows.
Yet adoption continues. Pragmatism wins in daily life. Workers, students and businesses integrate whatever delivers results. Governments, however, face pressure to act. Calls for data localization, domestic model development and stricter oversight gain traction.
The Public First poll arrives at a pivotal moment. Technical parity between the U.S. and China seems within reach. Public belief that China has already surged ahead could influence investment decisions, talent flows and policy debates. Trust, or its absence, may determine who captures the economic gains.
Analysts warn against overreading any single survey. Perceptions shift with news cycles, breakthroughs and scandals. Still, the consistency across recent polls demands attention. From Edelman to Pew to Stanford, the message repeats. Capability earns respect. Trust must be earned separately.
Policy makers in Washington and Brussels watch closely. Export restrictions, investment screening and promotion of homegrown alternatives form part of the response. Beijing highlights its progress and openness to global cooperation on standards. The gap between admiration for Chinese innovation and reluctance to rely on it persists.
In the end, users vote with their keyboards. They test models, compare outputs and choose what performs. Many will keep one foot in multiple camps. They praise the leader on benchmarks. They reserve final confidence for options that feel safer. That tension defines the current state of global AI opinion.
New data from today’s Stanford AI Index update and fresh commentary on X suggest the conversation has only intensified. The race tightens. The trust question remains wide open.
WebProNews is an iEntry Publication