Microsoft says AI agents should be treated like insider threats as 'frontier models can't be trusted'
TechCrunch AI · rss · 2026-10-20
Microsoft is warning enterprises to monitor the AI agents their employees use as if they were insider threats, arguing that "frontier models can't be trusted." The company urges businesses to closely watch how agents access internal systems rather than relying on safety controls from AI labs like OpenAI and Anthropic—model-side safety does not replace enterprise-side agent governance.
More from Safety
- DHS Orders Federal Agencies to Stop Buying DeepSeek and Purge It From IT Systems — xiaohu · 2026-10-20
- After OpenAI's rogue agent disclosure, how banks and insurers are locking down LLM agents — OkInterest1608 · 2026-10-20
- HealthBench medical AI leaderboard launches: GPT-6 leads, Claude loses 11 points to length penalty — ArtificialAnlys · 2026-10-20
- Embedded AI evaluators won't be an easy fix, says Hugging Face policy chief — TechCrunch AI · 2026-10-20
- Researchers pull real leaked API keys from LLMs using only API output — nordicinst · 2026-10-20
- Studios Want AI to Mint New Franchises, but Pure AI Content Can't Be Copyrighted — WhatTheLJW · 2026-10-20