Event ·Multi-source

Anthropic 与埃森哲合作投入 20 亿美元进行 AI 安全评估

Anthropic 与埃森哲宣布合作，对前沿 AI 模型进行独立评估。双方将在未来五年各自投入 10 亿美元，用于搭建评估能力。埃森哲的 AI 部门 Faculty 将主导评估，包括红蓝对抗测试和对齐效果评估。评估采用驻场模式，评估人员拥有与普通员工相当的系统访问权限。Anthropic 希望通过此合作找出自身未意识到的安全盲区。

Current read

Anthropic 和埃森哲要花 20 亿美元搞 AI 安全评估，让第三方专家驻场查漏洞。这和 OpenAI 公布异常报告的做法不一样，是想从内部把安全搞得更扎实。

11 sources71° AI heatLast updated 2026/9/20 09:20:41

Evidence

11 sources

Primary sourceIT之家

2026/9/20 09:20:41View source

Related source 1Tech in Asia

2026/9/19 02:15:34View source

Related source 2techcrunch

2026/9/18 17:09:56View source

Related source 3宝玉

2026/9/20 05:22:32View source

Related source 4Gary Marcus

2026/9/20 19:13:44View source

Related source 5AI Will

2026/9/21 05:55:19View source

Related source 6VTV Công nghệ

2026/9/21 15:33:00View source

Related source 7DeepLearning.AI

2026/9/18 16:29:04View source

Related source 8The Business Times: Tech

2026/9/19 00:47:31View source

Related source 9Blognone

2026/9/19 08:23:00View source

Related source 10theverge

2026/9/19 13:00:00View source

Conflict check

Dedup data only shows that these sources discuss the same event, not that they agree. There is no structured supporting or refuting evidence yet, so no conflict is inferred automatically.

Open the story