Microsoft says AI agents are like insider threats as ‘frontier models can’t be trusted’
Microsoft wants enterprises to closely watch the agents that employees use with systems not controlled by AI labs like OpenAI and Anthropic.

Microsoft wants enterprises to closely watch the agents that employees use with systems not controlled by AI labs like OpenAI and Anthropic.
Key Takeaways
- •Latest on Microsoft says AI agents are like insider threats as ‘frontier models can’t be trusted’ — reported by TechCrunch AI for the news space.
- •This story was reported by TechCrunch AI, covering developments in the news space.
- •AI advancements continue to reshape industries — read the full article on TechCrunch AI for complete coverage.
📖 Continue reading the full article:
Read Full Article on TechCrunch AI →


