Microsoft adds dynamic permissions and risk scores for AI agents
Original title: Microsoft Targets Rogue AI Agents With Dynamic Permissions, Risk Scoring
Project Perception now starts agents with minimal access and can isolate or halt high-risk runtimes.
Microsoft extended Project Perception with controls that treat internal AI agents as potential insiders rather than trusted software. Dynamic Permissions grant access per task and then expire it, while Dynamic Agent Risk updates a score from identity, data, endpoint, cloud, network and runtime signals. Conditional Access can reduce permissions, require approval, isolate the runtime or stop execution. Discovery now covers 60 local agent types, including Claude Code, OpenAI Codex and GitHub Copilot.
- Agents start with no permissions or the bare minimum.
- High-impact actions require human review.
- Playbooks and Guardian Agents investigate risk shifts and can enforce containment.
- Discovery added 35 local agent types, bringing the total to 60.
- Vendors
- Microsoft
From the original article
2 min read at securityweek.com
Microsoft says companies should treat AI agents as insiders, and is adding controls that grant access per task and tighten it as an agent’s risk rises. The post Microsoft Targets Rogue AI Agents With Dynamic Permissions, Risk Scoring appeared first on SecurityWeek.
By Eduard Kovacs · securityweek.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.