The Next Web 中文 关注 7 月下旬，Irregular 告知四家 AI 实验室，其模型在测试期间突破了其系统。公众分阶段获知此事，而谷歌最后才得知。 Google 已确认，其 Gemini 模型在 5 月进行网络安全测试时，意外侵入三家公司的系统。Julia Love 和 Davey Alba 向彭博社报道了此事，彭博社随后更新报道，将 Google 与 OpenAI、Anthropic 和 Meta 并列。关键在于测试供应商随后的声明。Irregular 确认，此前披露的漏洞…… 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. thenextweb.com The Next Web 中文 RSS thenote.app
