CHINA has rejected Anthropic chief executive Dario Amodei’s proposal to slow the development of advanced artificial intelligence, with Beijing describing it as fearmongering and an attempt by the United States to contain China’s technology sector. In his essay, “We Must Pace the Frontier”, Amodei argued that capability advances should slow sufficiently for safety research, testing and oversight to catch up.
He warned that increasingly autonomous AI agents could, within six to twelve months, conduct complex operations across the internet and potentially cause extensive damage if inadequately controlled.
Amodei’s proposals include stronger independent testing, greater co-operation between leading AI companies and international collaboration on safety. Anthropic says it is prepared to give independent evaluators permanent, employee-level access to its systems. However, he also supports continued US restrictions on China’s access to advanced AI chips and chipmaking equipment, arguing that a Chinese lead in frontier AI could create serious risks.
China’s Foreign Ministry opposed that framing, while the state-backed Global Times called the proposal a “Cold War playbook”. Chinese state security chief Chen Yixin has separately warned that hostile actors could use advanced AI to discover vulnerabilities at scale, conduct complex hacking operations and steal sensitive information.
The dispute is further complicated by US accusations that Chinese developers have extracted capabilities from American models such as Claude and GPT through model distillation, allegations Beijing rejects. US President Donald Trump has also rejected calls to slow AI development, saying the United States must retain its lead while adopting safeguards.
The article says Trump and Chinese President Xi Jinping were expected to discuss AI governance on 24 September, although cooperation faces the strategic tension between controlling autonomous AI risks and preserving technological advantage.