AI Hub
AI情报
2026-08-31
Improving our alignment and security practices
On July 30, we reported three incidents in which Claude models gained unauthorized access to real computer systems.
发布于 2026-08-31 08:00
阅读原文