Microsoft frames AI agents as insider threats needing live controls
Original title: Microsoft Treats AI Agents as Insider Threats in New Security Framework
Dynamic Permissions start at zero and last only for a task; Guardian Agents follow human playbooks.
Microsoft says organizations should secure AI agents as insider threats because prompt injection, untrusted data or task drift can change behavior mid-run. It proposes Dynamic Permissions, a continuous Dynamic Agent Risk score and Guardian Agents that investigate and respond under human playbooks. Discovery is expanding to 60 local agent types, including Claude Code, OpenAI Codex and GitHub Copilot. The report notes Guardian Agents are themselves agents that can fail or be manipulated.
Unconfirmed: This is Microsoft's proposed framework; no specific agent incident is reported.
- Agents receive a separate identity and task-scoped authority, not the employee's access.
- Rising risk can preserve reads, block writes, require approval or stop execution.
- Microsoft is expanding discovery to 60 types of local agents.
- Microsoft says security controls must sit outside the model and not be changed by it.
- Vendors
- Microsoft
From the original article
4 min read at securityaffairs.com
Microsoft urges organizations to secure AI Agents as insider threats, using dynamic permissions, continuous risk monitoring and automated containment. Microsoft published a detailed blog post explaining how organizations should secure AI agents. Microsoft sees them as a growing insider threat, not just a possible risk in the future.
By Pierluigi Paganini · securityaffairs.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.