CrowdStrike unveils an agent kill switch at Fal.Con Europe on November 3 to shut down rogue AI agents
Part of the Falcon Guardian platform, it lets security teams stop rogue agents, prevent restarts and cut network access, with the CEO saying the industry previously lacked a way to shut agents down.
On November 3 local time, CNBC reported that CrowdStrike CEO George Kurtz said the company has shipped a new feature that can shut down runaway AI agents directly. CrowdStrike unveiled the AI kill switch at the Fal.Con Europe conference, letting security teams stop a misbehaving agent, prevent it from restarting and cut its network access. Kurtz said everyone has been building AI agents without a button to turn them off, and that this now exists and is critical to keeping agents running safely.
The AI kill switch is part of CrowdStrike's Falcon Guardian platform, which monitors AI agents running inside enterprise environments. The capability is generally available and targets the safety risks that have accompanied the rapid spread of agents.
In September, Anthropic CEO Dario Amodei called for slowing AI development over security concerns, and OpenAI CEO Sam Altman backed that position. Against that backdrop, bringing agents under unified monitoring with a forced-interruption capability has become an added layer in enterprise security architecture. As agents gain broader system permissions and network reach, a controllable interruption mechanism is shifting from an operational option to a hard requirement for compliance and audit.