AI Agents Act a Lot Like Malware. Here’s How to Contain the Risks.

ABSTRACT

HBR presents a governance-driven perspective on agentic AI risks, likening certain agent behaviors and threat models to malware. The core problem is organizational: how to contain agent risks with policies, controls, and oversight. The methodology is analogy-driven but policy-relevant, aiming to translate technical concerns into actionable risk containment steps.