Key points
- Microsoft treats autonomous AI agents as potential insider threats in new security tools.
- The Guardian Agent feature dynamically adjusts permissions based on real-time risk assessments.
- Project Perception coordinates red, blue, and green agents to manage these security risks.
Microsoft has introduced new security capabilities in its Microsoft 365 E7 suite and Project Perception system that treat autonomous AI agents as potential insider threats requiring continuous monitoring.
How it unfolded
- Microsoft announced new tools to help organizations manage security risks from autonomous AI agents acting as digital insiders, as Circuit Observer reported on 2026-10-20.
- Nadim Abdo, Corporate VP for Identity and Network Access, published a blog post arguing that agents need external oversight because they cannot trust themselves with their own security.
- Microsoft unveiled the Guardian Agent feature to monitor, constrain, and manage risks posed by these autonomous systems through dynamic permission adjustments.
Who is affected
Organizations deploying autonomous AI agents within Microsoft 365 E7 environments are the primary audience for these new controls. According to Microsoft, these agents require extensive permissions to function properly, which creates significant risk if not handled correctly. The company warns that "agentic insiders" require continuous AI-driven oversight beyond traditional security controls.
What happens next
Microsoft states it is extending Project Perception to secure "the agentic insider" using the new Guardian Agent. This system will coordinate specialized AI agents to analyze security data and identify vulnerabilities. The controls determine what an agent can access and whether it should continue operating, sitting outside the model to prevent tampering.
The background in plain language
Autonomous AI agents perform tasks independently, often needing broad access to company data. This access creates insider risk if the agent makes errors or is manipulated. Traditional security checks happen too slowly for machine-speed actions. Organizations now use separate monitoring agents to watch these workers, adjusting permissions dynamically based on behavior rather than static rules.
Background: Insider threats
You detect insider threats by correlating identity logs with data access patterns. Look for impossible travel, abnormal data exfiltration volumes, and privilege escalation outside standard workflows. Combine automated monitoring with manual review of high-risk accounts to catch anomalies that automated tools miss.
Read the full guide: Detecting Insider Threats: Log Signals, Blind Spots, and Hidden Costs
What to do and how to stay safe: Microsoft
- Review the permissions granted to all autonomous AI agents in your Microsoft 365 E7 environment.
- Verify that external oversight controls are enabled for any AI tools that access sensitive data.
- Monitor logs for unusual permission changes or automated actions that deviate from standard patterns.
- Ensure that security decisions are made by systems outside the AI agents themselves.
Step-by-step guide: Injection attacks examples: how prompt leaks break AI systems
General security guidance from the Circuit Observer newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What is a Guardian Agent?
A Guardian Agent is a new Microsoft security tool that monitors and constrains autonomous AI agents by dynamically adjusting their permissions based on risk levels.
Why does Microsoft treat AI agents as insider threats?
Microsoft states that AI agents can make mistakes or be manipulated, and their extensive permissions create risks that require continuous oversight beyond traditional controls.
How does Project Perception help with this?
Project Perception coordinates specialized red, blue, and green agents to analyze security data, identify vulnerabilities, and implement corrections for AI-related risks.
