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

companiesTechCrunch AIOct 19, 2026

Microsoft wants enterprises to closely watch the agents that employees use with systems not controlled by AI labs like OpenAI and Anthropic.

Impact: LowConfidence: MediumNovelty: Incremental
companiestechcrunch-aiagents

See the whole story

Read the original on TechCrunch AI

Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original source is authoritative.