For years, many influential figures in Silicon Valley have claimed that the United States is engaged in an AI arms race with China, warning of severe consequences if the U.S. falls behind. Recently, some of these executives have shifted their focus, suggesting that the two nations need to collaborate to mitigate the potential dangers of AI. Anthropic CEO Dario Amodei stated in a blog post published on September 16, 2026, that global pacing in AI development will require cooperation with China, which he describes as having the most advanced AI capabilities.
However, the Chinese government and its AI companies have expressed skepticism regarding Amodei's proposal, viewing it as an unfavorable deal. Kevin Xu, founder of Interconnected Capital and a former White House staffer, noted that the AI safety community in China shares concerns similar to those in the U.S., but they believe that the U.S. approach is aimed at subjugating China rather than fostering genuine collaboration.
Amodei emphasized that a key aspect of slowing down AI development should involve ensuring that democracies maintain a technological lead over autocracies like China. He suggested measures such as cracking down on chip smuggling and preventing unauthorized model distillation by Chinese firms. Alvin Wang Graylin, a senior fellow at the Asia Society Policy Institute, remarked that Amodei's proposal appears inauthentic to Chinese audiences because it conditions cooperation on maintaining a lead over China.
Amodei also posited that weakening China's AI industry could increase U.S. leverage in negotiations, a point he later acknowledged was unclear. Chinese officials and AI researchers have not dismissed the risks associated with AI but have expressed resentment towards being viewed as both a partner and a competitor that needs to be constrained.
Guo Jiakun, a spokesperson for China’s Ministry of Foreign Affairs, criticized Amodei's proposal as fear-mongering and detrimental to global AI cooperation. The Chinese leadership is primarily concerned about domestic stability and the potential threats AI poses to the Communist Party's authority, rather than existential risks to humanity.
Michelle Nie, a visiting fellow at the Center for a New American Security, noted that Chinese policymakers are skeptical of U.S. narratives about AI risks, viewing them as self-serving. A researcher from DeepSeek expressed a belief that AI should be accessible to all, criticizing companies like Anthropic and OpenAI for their control over advanced AI technologies.
Despite the skepticism, there are indications that U.S.-China cooperation on AI safety is not impossible. Preliminary negotiations are reportedly underway ahead of a planned meeting between Chinese leader Xi Jinping and U.S. President Donald Trump. Both governments share concerns about cybersecurity threats and the misuse of AI technologies.
Nie suggested that there are individuals on both sides who genuinely care about AI safety and could engage in constructive discussions. However, others caution that rhetoric from U.S. tech leaders about weakening China could undermine trust necessary for diplomatic talks. Kristy Loke from MATS Research warned that such statements may ultimately hinder efforts to address the risks associated with AI.