# Bloomberg：北京以 Anthropic 为试金石，回应美国主导的前沿 AI 安全规则

- 来源：Rohan Paul (@rohanpaul_ai)
- 发布时间：2026-09-01 09:10
- AIHOT 分数：63
- AIHOT 链接：https://aihot.virxact.com/items/cmthzajex05vqrofqjg89ggts
- 原文链接：https://x.com/rohanpaul_ai/status/2094593749120901314

## AI 摘要

Bloomberg 报道，北京的政策信号显示美国 AI 公司须面对华盛顿在海外提出的规则，Anthropic 成为中方挑战美国定义的前沿 AI 安全规则的试例。中方希望危险能力由双方共同定义和管控，而普通中国 AI 竞争留在限制之外；相关批评通过 CCTV 关联账号而非正式政府声明发出，属于强烈政策信号而非已宣布的谈判立场。

## 正文

Bloomberg: Beijing's policy signal says American AI companies must face the rules Washington proposes abroad.

Anthropic has become China's test case for challenging U.S.-defined frontier AI safety rules. China wants dangerous capabilities jointly defined and controlled, while ordinary Chinese AI competition should remain outside those restrictions.

The Anthropic criticism came through a CCTV-affiliated account rather than a formal Chinese government announcement. i.e. its a strong policy signal rather than an announced negotiating position.