Microsoft says AI agents are like insider threats as 'frontier models can't be trusted' | TechCrunch

The short version

Microsoft warns enterprises of AI agent risks, likening them to insider threats, and introduces Guardian Agents for enhanced monitoring.

Key takeaways

• Microsoft emphasizes AI agents can pose security risks.
• Guardian Agents will monitor enterprise AI usage.
• AI agents should be controlled independently from their providers.
• Nadella stresses responsibility for AI outcomes must remain with enterprises.

Overall sentiment

The article conveys concerns about AI risks, expressing distrust in their management.

AI classification of the article’s tone. Coverage is focused solely on Microsoft's introduction of security measures for AI agents.

Sources cited by the article

• introduced new security features ↗
• continues to warn enterprises about the dangers ↗
• Guardian Agents ↗
• garnered the support of more than 100 other tech companies. ↗
• Surface Laptop Ultra PCs built for agents ↗
• AI ↗
• Enterprise ↗
• Microsoft ↗

These references were linked or listed by the publisher, then selected for relevance. They are distinct from our original-source citation.