loading
loading
China is pushing back on Anthropic CEO Dario Amodei’s attempt to rewrite the rules of artificial intelligence. Earlier this month, Amodei called for the global AI industry to slow the development of its most powerful models, arguing that the pause would buy some time to make them safer. The proposal, he wrote in a blog post, would “require cooperation with China, the autocratic country with by far the most advanced AI capabilities.” Amodei went on to outline measures to slow China’s AI development “to give us the breathing room we need in order to pace effectively.” He called for an end to sales of powerful AI chips or semiconductor manufacturing equipment to China, a crackdown on chip smuggling, and restrictions on remote access to data centers outside China, among other measures. Amodei won unusual support from U.S. competitors, including OpenAI CEO Sam Altman and xAI founder Elon Musk. Meta CEO Mark Zuckerberg and President Donald Trump rejected the idea, instead calling for rapid, unfettered AI growth. China’s answer has been a clear no. Officials, state media, and engineers have pushed back against a proposal they see as another move to preserve the U.S.’s technological lead. “I don’t see any reason why China would cooperate with [Amodei’s] proposal and be held down more than it already is,” Alvin Wang Graylin, a senior fellow on technology at the Asia Society Policy Institute’s Center for China Analysis, told Rest of World. But that does not mean China would never agree to pace AI development, Graylin said. It means “not agreeing to abide by [Amodei’s] terms as is.” Trump is scheduled to meet Chinese President Xi Jinping in Washington on September 24, with Washington expected to raise the “shared risks” from AI with Beijing. Cold War playbook? Beijing wasted little time responding to Amodei. A day after the blog post was published, Chinese state-backed Global Times ran an editorial titled “Targeting China’s AI: U.S. ‘tech right’ unfolds Cold War playbook.” It argued that Amodei was proposing restrictions on chips and Chinese AI companies to preserve American technological dominance. On September 11, Chinese Foreign Ministry spokesperson Guo Jiakun described Amodei’s call as “fear-mongering, confrontation and vicious competition,” and said it would undermine global AI governance. But China has its own increasingly expansive definition of AI safety. China’s national cybersecurity standards body released a third edition of its Safety Governance Framework on September 14. The document warns that AI could autonomously learn, optimize itself, and undergo “recursive self-improvement” at a speed or in a direction that exceeds human expectations and control. This concern echoes the kind of dystopian scenario that former Anthropic researcher Jacob Coxon warned about earlier this month. “I do not think China and the U.S. see AI risks in fundamentally different ways,” Zilan Qian, a London-based research scholar at think tank GovAI, told Rest of World. “But their socioeconomic conditions and political institutions make some risks more visible and some voices more prominent in each country.” Is China worried about AI dystopia? Chinese policymakers are concerned about a range of AI risks, including its influence on the information ecosystem, people forming romantic relationships with chatbots, its potential effects on labor, and its use by malicious actors to commit cybercrime, Qian said. “Compared with U.S. frontier labs, there is less discussion in China of frontier AI risks, such as humans losing control of AI or AI automating AI research,” Qian said. “However, these risks are increasingly being acknowledged by Chinese labs and in policy documents.” Days before Amodei’s essay, Chinese State Security Minister Chen Yixin warned that advanced U.S. models such as Anthropic’s Mythos and OpenAI’s GPT-5.5-Cyber could pose serious risks to China’s critical information infrastructure, among other threats AI could pose. Chen called for tighter government oversight, greater technological self-reliance and stronger defenses against AI-enabled attacks and influence operations. “Modern warfare has entered an era increasingly shaped by artificial intelligence,” Chen wrote in a state-run magazine, China Cyberspace, on September 13. “Whoever can use algorithms to achieve more refined sensing, more accurate decision-making, and more precise strikes will gain the initiative on the battlefield.” Who defines the game? China’s leading AI companies have largely stayed silent on Amodei’s proposal. Some of the engineers building their models, however, have been less restrained. Tao He, head of R&D at MiniMax Code, an AI agent app at the Chinese frontier lab, was among the few engineers who have responded publicly. “Oh, come on. Yesterday, the mission was AGI. Today, we’re told to slow down,” He wrote the day after Amodei published his essay. “Why do you always get to define the game — and change the rules?” He did not respond to Rest of World’s request for comment. DeepSeek engineer Liu Shengyu, who worked on the Chinese AI company’s latest model, described his own version of AI dystopia in a personal essay published on his WeChat account. Liu imagines a world in which a tiny number of technology companies control society’s most powerful AI — and with it, an overwhelming share of economic and technological power. Those with access to frontier intelligence, he wrote, would become increasingly difficult for everyone else to catch, creating a vicious cycle of technological and social inequality. He called for making powerful AI cheap and widely available. “I don’t trust Anthropic or OpenAI to do that. In particular, I do not want Anthropic to control the world’s most advanced AI or AGI,” Liu wrote. For Chinese engineers already building under U.S. restrictions on advanced chips, calls to slow down can carry a different meaning, particularly when their models are still behind America’s frontier. Huawei rotating chairman Eric Xu said last week that leading U.S. labs may be encountering risks that Chinese developers cannot yet perceive. Rather than slow down, he suggested Chinese developers may first need to catch up. “I think maybe AI model providers in China may need to speed up their pace to the level that they could also feel the risks from AI development,” Xu said.
document fetched 2026-09-24 02:52 · canonical https://restofworld.org/2026/china-rejects-dario-amodei-anthropic-ai-slowdown-trump-xi