An Essay, and a Sharp Response
Anthropic CEO Dario Amodei published an essay — reported to be titled "We Must Pace the Frontier" — arguing that AI companies should deliberately slow the pace of frontier model capability gains while independent evaluators verify safety practices. China's Ministry of Foreign Affairs pushed back within days, framing the proposal as thinly veiled containment rather than genuine safety cooperation.
The exchange is the latest flashpoint in the broader US-China dynamic around AI development, export controls, and who gets to set the pace of the technology's advance.
What Amodei Argued
According to reporting on the essay, Amodei made a few connected points:
- A "Chinese lead in AI would pose grave danger for the United States and the world," he wrote, and argued the US should maintain existing restrictions on selling cutting-edge AI chips and chipmaking equipment to China to preserve America's technical edge.
- He acknowledged what he called the "toughest dilemma" in his own proposal: asking frontier labs to slow down only works if rivals — namely China — do the same, and there's no mechanism forcing that.
- Speaking to CBS News, Amodei reportedly said the "more long-term thing would be working together to put a speed limit on the rate of AI progress," while conceding "I think that's going to be very difficult because the incentives to pull ahead and the military advantage that you get from that are so large."
The essay drew public support from some rivals: Sam Altman posted agreement with the "pace the frontier" framing on X, Demis Hassabis (chair of Google DeepMind) said the essay "points towards the right path forward," and Elon Musk also welcomed the proposal.
Beijing's Response
China's Ministry of Foreign Affairs addressed the essay through spokesperson Guo Jiakun at a regular press briefing, saying that all parties should work together on AI rather than treat it as a competitive threat. He was quoted as saying:
"Fearmongering, confrontation and vicious competition will only disrupt the process of global AI governance, which serves no one's interest."
The state-backed Global Times went further, calling Amodei's proposal a "Cold War playbook" aimed at restraining China's technological rise. In an editorial, it characterized the plan's "true agenda" as an attempt to "curb China's AI development through technological barriers and regulatory monopolies, uphold Washington's monopolistic hegemony in cutting-edge technology, and exclude China from the global AI governance system" — describing it as a "silent AI Cold War" that was "hypocritical and short-sighted."
Separately, China's State Security Minister Chen Yixin described AI as "a new arena for strategic rivalry among major powers" that could threaten political security, critical infrastructure, and ideological stability — remarks reported alongside, but not directly tied to, the response to Amodei's essay.
China's Commerce Ministry also weighed in on a related front, dismissing US claims about Chinese firms "distilling" capabilities from Western AI models as groundless, arguing that model distillation is a technique used industrywide, and accusing the US of pursuing a "monopoly" over the AI industry through export controls.
The Backdrop
A few pieces of context help explain the timing:
- The essay landed a few days after the FBI, NSA, and CISA issued a joint cybersecurity advisory alleging that Chinese AI developers had engaged in "aggressive, malicious" efforts to extract, or "distill," capabilities from advanced US models including Anthropic's Claude and OpenAI's GPT models.
- Presidents Trump and Xi are reportedly set to meet on September 24, with AI governance expected to be among the topics discussed.
- Trump has separately downplayed calls to slow US AI development, saying the US is leading China on AI and that "whoever wins AI, wins."
Why This Matters
Stripped of the rhetoric on both sides, the disagreement is really about two different theories of how AI risk gets managed: Amodei's essay argues for deliberately capping the pace of capability growth (paired with continued US chip export restrictions to preserve a lead), while Beijing's response frames any unilateral "pacing" proposal — especially one bundled with export controls — as a US attempt to lock in technological dominance rather than a genuine safety measure. Neither government has proposed a concrete, mutually verifiable mechanism for actually slowing development, so for now this remains a public argument about framing rather than a policy with teeth.
Sources
- SecurityWeek — Beijing Hits Back at Anthropic CEO's Call to Curb China's AI Development
- NPR — China Hits Back at AI Development Warning
- CNBC — Anthropic's Amodei Says China Presents 'Toughest Dilemma' for His Proposed AI Slowdown