CrowdStrike has introduced a tool that lets security teams stop AI agents that behave unexpectedly. Announced Tuesday at the company’s Fal.Con Europe conference, the AI Kill Switch can terminate an agent, prevent it from restarting and sever its network access, CEO George Kurtz said.
The tool is part of CrowdStrike’s Falcon Guardian platform, which monitors AI agents used within an organization. It is available immediately and is intended to give businesses more control over systems as they adopt autonomous AI tools.
“Everyone is building AI agents, but they never had the off button, and now they do,” Kurtz told CNBC’s Jim Cramer. He said the capability is important for ensuring agents operate safely and can be stopped if they go beyond their guardrails.
Kurtz described the launch as an expansion of cybersecurity’s role: alongside defending companies against external threats, providers must help ensure AI systems behave as intended inside organizations. He called safe deployment of agents a significant market opportunity for CrowdStrike.
The CEO also pointed to customer demand for security measures around AI adoption, citing CrowdStrike’s partnership with Anthropic to protect its Claude Cowork product. He said enterprise customers asked for CrowdStrike’s technology to be integrated before deploying Anthropic’s offering.
Sources: CNBC Markets.