Microsoft launches Guardian Agents to monitor enterprise AI agents

Microsoft introduced Guardian Agents, part of Project Perception, to monitor AI agents used by enterprises. CEO Nadella has warned that frontier models can't be trusted and that companies must use independent systems, not controlled by model providers, to oversee agent behavior.

• AI worms and major incident fears drive safety calls at Dublin summit2026-10-14
• Reports of OpenAI model misalignment grow, prompting calls for stronger guardrails2026-10-14
• Microsoft President Says AI Less Regulated Than Milk2026-10-12
• Google Launches 'Universal' Gemini AI Agent for Workplace2026-10-08
• Australian inquiry demands OpenAI explain AI data access2026-10-05