Microsoft says AI agents are like insider threats as ‘frontier models can’t be trusted’

2026-10-20来源 TechCrunchEnglish
Microsoft wants enterprises to closely watch the agents employees use with systems not controlled by AI labs like OpenAI and Anthropic.

原文：https://www.a8q9.cn/en/article/microsoft-says-ai-agents-are-like-insider-threats-as-frontier-models-can-t-be-tr-a16a35.html

相关文章