CrowdStrike has introduced an AI security feature that allows enterprise security teams to stop rogue or misbehaving AI agents, prevent them from restarting and restrict their network access. The new “AI Kill Switch” was announced on November 3 at the company’s Fal.Con Europe conference. The feature is part of CrowdStrike’s Falcon Guardian platform, which monitors AI agents operating within organisations. It is designed to give security teams greater control when an AI agent acts outside its intended instructions or performs unauthorised activities. According to CrowdStrike, the kill switch allows authorised security personnel to terminate an agent, block it from launching again on a targeted system and cut off its enterprise network access. The company says the action is initiated by a human administrator, with the response recorded in an audit trail. The feature is available to Falcon Guardian customers. CrowdStrike says its existing platform can identify AI agents, enforce policies on their activities and detect attacks targeting them. The announcement comes as businesses increasingly deploy AI agents capable of interacting with applications, systems and enterprise resources. Unlike conventional AI assistants, these agents can carry out multiple steps and execute actions with varying levels of access, creating additional security concerns. CrowdStrike’s move reflects a growing focus on AI safety and enterprise cybersecurity as organisations adopt autonomous systems. The company is positioning the kill switch as a way for businesses to retain human oversight while using AI agents. The feature is intended to help security teams respond when agents behave unexpectedly, although its effectiveness in preventing real-world incidents has not been independently established.

AdvertisementThe Puranic