CrowdStrike Falcon Guardian: AI Agent Security Launched

CrowdStrike has launched Falcon Guardian, a new AI detection and response product. This product is designed to provide runtime security for enterprise AI agents. Guardian covers a wide range of elements including data, models, prompts, agents, identities, infrastructure, and interactions. It also extends across endpoints, cloud environments, SaaS applications, and browsers. CrowdStrike CEO George Kurtz stated that AI demands the same approach to security that pioneered endpoint detection and response. He emphasized that AI changes the speed of attacks, and Falcon Guardian turns policy into protection. The Falcon sensor discovers AI agents on Windows and macOS, creating an inventory and identifying who deployed them. Agent Runtime Visibility links agent behavior to endpoint data, tracing user prompts and actions. Agent Access Controls define which AI agents can run and block unauthorized ones. The Runtime Detection and Response tool detects attacks and malicious agent behavior, reconstructing execution chains and containing threats. AI Gateway will provide a central control point for enterprise AI traffic. Falcon Complete for Guardian offers 24/7 expert-led detection and response for AI agents. Falcon Adversary OverWatch for Guardian extends managed threat hunting to AI agent activity. Guardian also feeds AI agent data into Falcon Next-Gen SIEM for correlation across various security domains. This new offering aims to protect organizations from the evolving threats posed by AI agents.