FrontRow

AI Agents Among Us: Identify, Understand, and Control Every AI Agent Action

BeyondTrust argues AI agent risk is a privileges problem and shows how it controls what agents can do.

19 November 2026
10:00 – 10:45 GMT · 45 min
Online · Registration closed
WebinarArtificial Intelligence & CopilotCybersecurityIdentity & Access ManagementFree

Coming up instead

About this session

Would you give an AI agent the same privileges as your users? BeyondTrust says most organisations it speaks to answer no, and this 45-minute session is built on that. Its argument is that the risk from AI agents is mostly about what they're permitted to do, not about spotting malicious behaviour after the fact.

The session covers defining up front which actions an agent can take on an endpoint, telling the human and the agent apart in the audit trail, and how its Phantom Labs researchers used an AI worm to design the approach.

This is a product launch for BeyondTrust AI Agent Security, so treat it as a vendor session. The thinking on privileges for agents is still worth hearing if you're rolling AI tools out to staff devices.

Who it’s for

Security and identity leads, and IT managers, deciding how much access AI agents should have on staff devices.

Keep going

Something wrong with this listing? Tell us and we will put it right.