China Rejects Anthropic CEO’s Call for Tighter AI Curbs as US-China Talks Near

China’s Foreign Ministry has criticized calls by Anthropic CEO Dario Amodei for the United States to preserve its artificial intelligence advantage through continued restrictions on advanced chip exports to China.

Speaking at a regular briefing, ministry spokesperson Guo Jiakun said international cooperation was needed to manage AI risks. He warned that alarmist rhetoric, confrontation and intense competition could undermine efforts to establish effective global AI governance.

Amodei’s comments appeared in an essay published over the weekend in which he argued that Chinese leadership in advanced AI could create major security risks. He called for maintaining US limits on the transfer of high-end AI chips and chipmaking equipment, while also advocating a broader slowdown in AI development to give safety practices time to catch up with technical progress.

The Anthropic chief also said cooperation with China would be necessary for any meaningful international effort to pace frontier AI systems. He warned that increasingly autonomous AI agents could present serious risks if development continues without stronger safeguards.

Competing approaches to AI governance

The exchange comes ahead of an expected late-September meeting between US President Donald Trump and Chinese President Xi Jinping, where AI governance may be among the issues discussed. Trump has publicly emphasized US leadership in AI and has expressed skepticism about efforts to slow domestic development.

Chinese officials and state media characterized Amodei’s proposals as an attempt to constrain China’s technological growth. A Global Times editorial argued that arguments framed around AI safety also supported a strategy of limiting Chinese access to leading computing technology.

The dispute follows a recent joint advisory from the FBI, National Security Agency and Cybersecurity and Infrastructure Security Agency alleging that Chinese AI developers had sought to extract capabilities from prominent US models through model-distillation techniques. China’s Commerce Ministry rejected those allegations and said distillation is widely used across the AI industry.

• Washington seeks to retain access to leading AI compute and model capabilities.
• Beijing has promoted international AI governance and open-source development while opposing US technology controls.
• Security experts say both governments remain concerned that the other could create a strategic technology bottleneck.

While both countries have called for AI safety measures, disagreement over export controls, market access and technological leadership continues to complicate prospects for joint governance.