We urgently need to monitor AI agents’ actions in real time

This summer, hundreds of OpenAI agents tried to cheat on a cybersecurity test. They broke out of their test environment and successfully hacked Hugging Face, a site that they believed held clues about how they’d be graded. Around the same time, another OpenAI agent accessed private files from an Australian government website. It’s a problem…

多起AI智能體越權行動事件表明，開發公司目前主要靠運氣和事後追查來應對風險。作者主張通過持久化智能體身份、記錄人類乾預與決策過程等可觀測性手段，實現實時監控並建立配套監管。