Translated by Claude, edited by Sinocism.com

Source: Global Times New Media | September 13, 2026

On September 12, a proposal that appeared to call for ending destructive competition in the AI industry quietly stripped away the disguise of some people in the US technology sector.

9月12日，一份看似倡导AI行业“反内卷”的倡议，悄然撕下了美国科技圈一些人的伪装。

That day, Dario Amodei, chief executive officer of US artificial-intelligence company Anthropic, published an article calling for “controlling the pace of frontier AI development” to avoid the risk of technology spinning out of control. The proposal quickly won joint support from US AI heavyweights including Elon Musk and Sam Altman.

美国人工智能企业Anthropic首席执行官（CEO）阿莫迪当天公开发文，呼吁“我们必须控制前沿（AI）发展节奏”，以规避技术失控风险。这一倡议迅速得到马斯克、奥尔特曼等美国AI圈大佬的联合支持。

On the surface, this was a “rational appeal” focused on global AI safety. But a close reading exposes the trick: the article conceals numerous provisions aimed at containing China and is, in substance, a “Cold War script” for the AI field. Amodei claimed that democracies “absolutely cannot be naive” in dealing with China. To prevent China from surpassing its rivals while the United States “slows the pace of AI research and development,” he proposed first choking off China to widen the gap, then conditionally seeking negotiations with China on slowing down together. His approach included three specific restrictions and containment measures: a strict ban on selling high-performance AI chips and semiconductor manufacturing equipment to China; a crackdown on so-called model “distillation”; and preventing the theft or leakage through hacking of the core weights of large models.

表面看，这是着眼于全球AI安全的一次“理性发声”，可细读全文便能发现其中的猫腻：文章暗藏大量针对中国的遏制条款，实质上是一出AI领域的“冷战剧本”。阿莫迪声称，民主国家与中国打交道“绝对不能天真”。为防着中国借美方“放缓AI研发节奏超越对手”，他提出先对中国“卡脖子”拉开差距，再有条件地寻求与中国谈判共同减速的思路，其中包括三大具体限制和遏制措施：严禁向中国出售高性能AI芯片和半导体制造设备、打击所谓模型“蒸馏”行为以及防止大模型的核心权重被窃取或通过黑客手段外泄。

The rapid iteration of AI technology in recent years has indeed brought security risks, ethical concerns and the danger of technological loss of control—a global problem confronting all humanity. But the core of Amodei’s proposal completely departs from the original aspiration of joint governance of science and technology and politicizes a technological issue. This “quiet AI Cold War” is hypocritical and short-sighted. Its true purpose is to use technological barriers and a monopoly over rules to contain the development of Chinese AI, preserve Washington’s monopoly and hegemony in frontier technology, and exclude China from the global AI governance system. In essence, it is of a piece with Washington’s so-called “Pax Silica” initiative, which excludes China.

近年来AI技术飞速迭代的确带来了安全风险、伦理隐患、技术失控危机，这是全人类共同面临的全球性难题。可阿莫迪倡议的核心，完全背离了“科技共治”的初心，将技术问题政治化。这场“静悄悄的AI冷战”，虚伪且短视。其真正目的，是试图用技术壁垒、规则垄断遏制中国AI发展，护持华盛顿在前沿科技领域的垄断霸权，同时将中国排除在全球AI治理体系之外，本质上与华盛顿排斥中国的所谓“硅和平”倡议一脉相承。

Amodei’s attempt to discuss global AI safety while “containing China” is both a completely mistaken judgment and a fundamentally unrealizable fantasy. As one of the largest and most open suppliers in the global AI ecosystem, China’s AI capabilities have long since ceased to be a variable that can be “excluded.” Forcibly severing the global AI industrial and innovation chains and excluding China and Chinese companies would only greatly increase the trial-and-error costs and risks of loss of control in global AI development. Such “control of the pace” would itself lose all meaning.

阿莫迪试图在“遏制中国”的同时谈论全球AI安全，既是完全错误的判断，也是根本无法实现的妄想。作为全球AI生态中体量最大、开放程度最高的供给方之一，中国的AI能力早已不是一个可以被“排除在外”的变量。如果强行割裂全球AI产业链、创新链，将中国和中国企业排除在外，只会大幅提升全球AI发展的试错成本与失控隐患，这样的“控制节奏”本身已经丧失意义。