Does China agree with US AI threat warnings?

Silicon Valley’s mood has turned increasingly serious in recent weeks, with several technology executives warning that the most advanced artificial intelligence systems could eventually become difficult for humans to control.

Silicon Valley’s mood has turned increasingly serious in recent weeks, with several technology executives warning that the most advanced artificial intelligence systems could eventually become difficult for humans to control.

Against this backdrop, Anthropic CEO Dario Amodei called last Saturday for “worldwide pacing of the frontier,” arguing that such an effort would “require cooperation with China.” Beijing, however, has pushed back against proposals that it sees as attempting to slow China’s AI development on terms set by the United States.

Lian Jye Su, an analyst at technology research group Omdia, said China is “definitely not receptive” to a slowdown shaped by US technology companies. He described AI as a “core sovereignty capability” that Beijing would have little interest in restricting after years of efforts to narrow the gap with US frontier models.

Amodei has also argued that the United States should maintain “democracies’ AI lead over autocracies” and has continued to support export controls on advanced AI chips and related technology. Some analysts view such restrictions as part of a broader US strategy to limit China’s access to the computing resources needed to develop frontier AI systems.

China’s Foreign Ministry has rejected the idea of slowing AI development through confrontation. Responding to questions about Amodei’s proposals, spokesperson Guo Jiakun said that “fearmongering, confrontation and vicious competition” would disrupt efforts toward global AI governance.

At the same time, China has been developing its own AI safety and governance framework. The country recently released an updated framework that addresses emerging risks associated with AI agents and other advanced systems, indicating that Beijing is also focusing on AI safety and governance, even as it rejects calls for restrictions that it views as limiting its technological development.

Gabriel Wagner, a researcher at Concordia University, said there may be more common ground between the US and China on AI safety than their political rhetoric suggests.

“There is more common ground than people might assume between what Amodei is saying and what the Chinese government is saying,” Wagner said.

China has also expanded its regulatory approach to AI in recent years, including rules governing algorithms and generative AI services. More recently, regulators have placed increasing attention on how AI systems operate, particularly as increasingly autonomous AI agents become capable of taking actions with limited human intervention.

Su argued that China’s centralized control over infrastructure and commercial activity could give Beijing advantages in responding to a major AI-related emergency. However, other analysts have cautioned that neither country appears fully prepared for a serious loss-of-control incident.

China’s latest Five-Year Plan also places AI among the country’s major areas of technological and economic development.

Meanwhile, President Donald Trump has dismissed fears about AI posing an existential threat, calling concerns about rogue AI a “hoax.” Trump is scheduled to meet Chinese President Xi Jinping in Washington on September 24, with AI governance expected to be among the issues discussed. The United States and China have already begun discussions on mechanisms for communicating about AI-related national-security risks, although significant differences remain over technology restrictions and the pace of AI development.