行业多源确认
Mustafa Suleyman 谈 Anthropic 离职风波：AI 自我改代码是安全审计难点
精选理由
Microsoft AI 的 CEO 亲自回应 Anthropic 那场辞职事件，直接点出 AI 改自己代码难审计这个安全痛点，观点挺少见。
Microsoft AI CEO Mustafa Suleyman 在播客节目《The Rest Is Politics: Leading》中表示，几周前 Anthropic 的大规模辞职与递归自我改进风险有关。他指出，当 AI 在越来越少人类监督的情况下修改自身代码时，会产生重大安全隐患。他同时承认，这类行为很难进行审计。
原文 · rohanpaul_ai
Mustafa Suleyman (CEO of Microsoft AI) links the major Anthropic resignation to recursive self-improvement risk
"And so how and when an AI modifies its own code with less and less human in the loop directing and scrutinizing that, that's where there is a big kind of safety risk, which I think is what triggered the big resignation from Anthropic a few weeks ago.
That's actually a pretty hard thing to audit."
---- Full video on "The Rest Is Politics: Leading" YouTube channel, (link in comment)