Beijing Just Called Anthropic’s CEO AI ‘Disarmament Talk’ a Cold War Trick
Chinese state media branded Amodei’s proposal as a disguised geopolitical attack, while President Trump dismissed it as fearmongering.
Anthropic CEO Dario Amodei published a lengthy essay on X Saturday urging AI companies to slow model development, writing that “we must slow the pace at which we improve the capabilities of AI models” to make “wise use of the time we gain” to manage emerging risks.
The three-step framework in his essay called for independent evaluators with employee-level access inside frontier AI labs, coordination among companies on shared safety standards, and international cooperation to manage AI risk at a geopolitical level.
Over the next 48 hours, Amodei’s appeal for restraint turned into the kind of great power standoff he wanted to avoid.
Rivals Endorsed the Essay, Then Amodei Escalated the Comparison
According to Reuters, OpenAI’s Sam Altman and xAI’s Elon Musk publicly backed Amodei’s framework, with Altman confirming OpenAI would adopt “independent evaluators with employee-like access” inside the company.
Amodei clarified that he was not calling for a halt to AI training or technical progress, but for deliberate pacing.
As Fox News reports, speaking to CBS Sunday Morning the next day, Amodei compared the U.S. China AI race to the Cold War standoff with the Soviet Union, warning both nations must manage the threat of “potential destruction” from each other.
He also told Fox News that arms-limitation-style disarmament talks are “an appropriate lens” for addressing the issue geopolitically.
Amodei acknowledged the risk of China and other rivals continuing to advance while the U.S. holds back, calling it the “toughest dilemma,” per Reuters.
He also urged Washington to further tighten chip export controls on China and crack down on Chinese labs allegedly distilling capabilities from American models.
Beijing Read the Essay as a Geopolitical Move, Not a Safety One
China’s response arrived fast and pointed, as Reuters details.
The state-backed Global Times called Amodei’s essay a “Cold War playbook” aimed at China, arguing its real goal was to curb China’s AI progress “through technological barriers and regulatory monopolies” rather than genuine safety.
Beijing’s view makes sense, given that Amodei’s idea is to keep America ahead in tech, making his proposal look more like a strategy to stay ahead rather than a call for safety.
Building on his August warnings that Congress would regulate AI “out of business,” President Trump rejected the proposal Sunday.
He brushed off calls from “very negative forces” to pace technological advance, and declared, “We’re leading China in AI… whoever wins AI wins.”
His position leaves Amodei’s call for coordinated restraint with little political support from the country he’s asking to lead it.
An Appeal for Cooperation That Nobody Wanted to Cooperate On
The key takeaway isn’t whether Amodei’s framework is technically sound, but how instantly an apolitical safety proposal was framed as a geopolitical move by every major player involved.
Amodei called for independent verification and international dialogue; Washington wants to win outright; Beijing sees the proposal as containment disguised as caution.
Taken together, it is very possible to interpret that if you frame AI safety like a Cold War treaty, people will treat it like one.
Ahead of upcoming U.S.-China summit talks, Amodei wanted a peaceful agreement, but his call to pace AI progress ended up feeding right into the deepening rivalry between the two nations.
Source: China state newspaper blasts Anthropic’s calls to slow AI as ‘Cold War’ tactic
