Fable 5 回归：Anthropic 想给 AI 越狱定级
六月三十日，Anthropic 发布了一篇危机复盘式博客。按它的说法，Claude Fable 五会在七月一日恢复全球访问，Mythos 五也已恢复给部分美国组织使用。表面上是模型重新上线，真正的问题是：当强模型被发现可以绕过部分安全限制时，行业和政府该怎么判断严重程度。
0:00 / 5:27
Fable 5 回归：Anthropic 想给 AI 越狱定级
节目导览
Fable 5 恢复全球访问，Anthropic 同时提出一个给 AI 越狱风险定级的行业框架。本期解读这件事为什么不只是模型重新上线，而是前沿模型发布、政府测试和安全研究协作之间的新协调方式。
来源
- Anthropic｜Redeploying Fable 5
- Anthropic｜Statement on the US government directive to suspend access to Fable 5 and Mythos 5
- Anthropic｜Claude Fable 5 and Claude Mythos 5
- The White House｜Promoting Advanced Artificial Intelligence Innovation and Security
- NIST｜Center for AI Standards and Innovation
- FIRST｜Common Vulnerability Scoring System SIG
- Anthropic｜Expanding Project Glasswing
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.