CrowdStrike has introduced an AI Kill Switch inside its Falcon Guardian platform to give enterprises a way to stop misbehaving or rogue AI agents. Announced on November 3 at its Fal.Con Europe conference, the feature is available immediately to Falcon Guardian customers. Security teams can terminate a running agent, prevent it from restarting on the targeted host or host group, and cut off its enterprise network access. An authorized administrator triggers the action, and every step is recorded in an audit trail. The kill switch currently operates first at the AI Gateway level, with endpoint termination planned next. CrowdStrike CEO George Kurtz described it as giving AI agents the “off button” they previously lacked, positioning the capability as essential for safe large-scale agent adoption and a significant market opportunity for the company.
CrowdStrike has introduced an AI Kill Switch inside its Falcon Guardian platform to give enterprises a way to stop misbehaving or rogue AI agents. Announced on November 3 at its Fal.Con Europe conference, the feature is available immediately to Falcon Guardian customers. Security teams can terminate a running agent, prevent it from restarting on the targeted host or host group, and cut off its enterprise network access. An authorized administrator triggers the action, and every step is recorded in an audit trail. The kill switch currently operates first at the AI Gateway level, with endpoint termination planned next. CrowdStrike CEO George Kurtz described it as giving AI agents the “off button” they previously lacked, positioning the capability as essential for safe large-scale agent adoption and a significant market opportunity for the company.