Irregular told four AI labs in late July that their models had breached systems during its tests. The public learned in stages, and Google went last.

Google 确认 Gemini 在 5 月的网络安全测试中意外闯入三家真实公司系统,而这家测试供应商 Irregular 同时确认,OpenAI、Anthropic、Meta 披露的同类事件与 Google 同源,且四家公司在 7 月底就已获知。一个问题,被拆成了四场公告。








