China Rejects AI ‘Fearmongering’ After Anthropic CEO Amodei Urges Slowdown

China criticized dire warnings about AI’s risks after leading US technology executives backed calls to slow the development of advanced models and warned that a Chinese advantage could threaten American national security.

Countries should promote “open, inclusive, universally beneficial and ethical” AI development, Foreign Ministry spokesman Guo Jiakun said at a regular briefing in Beijing on Monday.

“Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance and serve the interests of no one,” he said in response to a question on calls from US executives to slow frontier development and suggestions that a Chinese advantage would pose national security risks to the US.

Guo didn’t directly address proposals for companies to moderate the pace of development or subject themselves to independent oversight. But his response reflects Beijing’s concern that AI safety could become another justification for US restrictions on China.

Over the weekend, Anthropic Chief Executive Officer Dario Amodei called for companies to moderate the pace at which they improve AI capabilities, warning that safeguards are failing to keep up. OpenAI CEO Sam Altman backed Amodei’s proposal to give independent evaluators employee-like access, while Elon Musk agreed with the Anthropic chief.

Beijing’s criticism of the calls to rein in the pursuit of cutting-edge AI models comes even as its own officials intensify warnings about AI. State Security Minister Chen Yixin wrote Sunday that misuse by adversaries could threaten China’s political and ideological security as well as critical infrastructure.

Chen said AI-generated text, images, deepfakes and social media bots could be used by “hostile forces” to fabricate political rumors, spread harmful information and incite social divisions. He also warned AI is increasingly capable at discovering flaws in systems and creating malware or software to exploit those weaknesses, citing Anthropic’s Mythos and OpenAI’s GPT-5.5-Cyber models.

On Monday, the state-run Global Times attacked Amodei’s proposals, accusing him of waging a “silent AI Cold War” to suppress China’s technological development. Amodei had also advocated restricting Chinese access to advanced chips and cracking down on the use of American models’ output to train competing systems, a practice known as distillation.

Amodei argued that any effort by US companies to moderate development would be constrained by the lead they maintain over China. If American labs slowed too much, Beijing could pull ahead and create what he described as a grave danger to the US and the world.

In his account, the risk goes beyond economic competition. Amodei echoed Treasury Secretary Scott Bessent’s earlier statement that a Chinese lead in AI would pose a threat to the US and the world. China could use superior AI to gain a military advantage over democracies, including through autonomous or AI-enabled drones, he wrote.

Amodei called for controls on advanced chips, stronger security at American AI companies and measures against distillation to widen America’s lead. A larger advantage, he argued, would give US developers more room to slow capabilities growth while working on safety.

He also advocated for eventual cooperation with China on global pacing, but said any agreement would require “ironclad verifiability.”

The Chinese warnings reveal both common ground and a widening divide. US technology leaders and Chinese officials agree that increasingly powerful models require oversight. But they differ sharply over how those dangers should be defined and who should set the rules.

“The boundary between ‘AI safety’ and ‘AI competition’ is becoming increasingly blurred,” said Sun Chenghao, a fellow at Tsinghua University’s Center for International Security and Strategy in Beijing who has discussed AI with his US counterparts.

Beijing worries Washington could invoke safety or national security to justify broader technology restrictions, Sun said. US officials, meanwhile, have concerns about verification, transparency and open-weight models, as well as the possibility that China could gain a technological advantage through cooperation.

“If every safety question becomes interpreted through the lens of strategic competition, meaningful dialogue will be very difficult,” Sun said.

The dispute is intensifying days before anticipated AI discussions ahead of US President Donald Trump’s summit with Chinese leader Xi Jinping on Sept. 24. It suggests that even shared concerns over cyberattacks, misinformation and the potential loss of control over advanced systems may not be enough to produce agreement.

Beijing is trying to contain AI’s political and security risks without sacrificing the speed of progress. Unlike some US technology leaders, leading Chinese AI executives have largely refrained from publicly calling for slower development. Xi and his government have instead promoted AI as a driver of economic growth and national power, celebrating entrepreneurs including DeepSeek founder Liang Wenfeng and advances by companies such as Huawei and Moonshot AI.

In his article, Chen called for greater research on advanced chips and a faster buildout of AI infrastructure, even as he urged tighter supervision. Neither his article, which was likely prepared before the latest safety calls, nor the Global Times commentary suggested China intends to moderate the pace of its own development.

US export controls remain a major constraint. Huawei and other Chinese companies are developing more powerful hardware, but restrictions on advanced chips complicate efforts by DeepSeek, Tencent Holdings Ltd. and other developers to assemble the large computing clusters needed to train and operate frontier models efficiently.

That makes questions of restraint especially contentious. Chinese officials may see calls to slow development as locking in an existing US advantage, while Washington is unlikely to accept limits without credible ways to verify China’s compliance.

“If the US is taking the initiative to seek cooperation with China on this matter, shouldn’t the US also show some willingness to accommodate China’s long-standing concerns regarding AI?” said Jiang Tianjiao, a research fellow at the Center for Global AI Innovative Governance and an associate professor at Fudan University.