Calls from Silicon Valley for the United States and China to slow frontier AI development are meeting resistance in Beijing because the proposals also seek to preserve a large US technological lead. Chinese officials and industry voices quoted by WIRED argue that cooperation cannot begin by weakening one side and treating it as an unequal partner.
Anthropic CEO Dario Amodei proposed tighter action against chip smuggling and unauthorized model distillation alongside eventual international pacing. China’s Foreign Ministry characterized that approach as confrontation and fear-mongering. Analysts said Chinese safety researchers share concerns about powerful systems, but Beijing’s priorities often center on disinformation, cyberattacks, social instability and threats to state control rather than the extinction scenarios common in some US labs.
The disagreement is therefore about both risk and bargaining terms. Open-weight developers in China also worry that a slowdown led by a few US companies could entrench their control over advanced models. Preliminary government talks reportedly remain possible ahead of a planned US visit by China’s leader, but the public reaction shows why a pact will be difficult: Washington-linked proposals frame technological advantage as leverage, while Chinese participants say meaningful safety rules require recognition as equals. Shared concern about AI does not automatically produce agreement on who may build it or set the limits.