CrowdStrike added an AI Kill Switch to Falcon Guardian, a platform designed to monitor corporate AI agents, track instruction sets, and block suspicious behavior at the gateway or endpoint.
CrowdStrike says that an administrator can trigger the shutdown when an agent exhibits suspicious behavior. Initially, the feature stops the agent at the point where it connects to the company's services. Soon, according to the company, it will be able to stop it on the device where it is running. It then prevents it from restarting on a specific computer or a selected group of computers and cuts off its access to the corporate network. Each action is triggered by an authorized administrator and logged.
Falcon Guardian detects and monitors AI agents in companies, whether they are approved or not. It monitors the instructions they follow and the files they can access, controls their access, and detects attacks using malicious instructions. It can also help investigate suspicious behavior.
Other companies have features for stopping such programs. Straiker has the Agentic Kill Switch, which can stop an agent, remove its access to tools, services, and data, or shut it down entirely. JetStream Security’s solution can disable individual agents without disrupting other AI tasks. Trust3 AI’s Agent DOS can be triggered by a human or automatically if an agent’s security score falls below a predetermined threshold.
↳
READ ALSO
HACKS.GR / RELATED
AI Unlocked
Athens becomes Wonderful's regional hub with the aim of quadrupling staff
↗
The source also cites Microsoft Entra Agent ID, Palo Alto Networks Prisma AIRS and Cisco AI Defense as solutions that can block agents. CrowdStrike founder and CEO George Kurtz said the new feature allows businesses to use AI agents while maintaining control.
>Stay connected with Hacks.gr
Set us as a preferred source on Google and find us on social media.
Preferred source
↗
end of article
Top comments (0)