CrowdStrike launches AI kill switch to control internal agents

CrowdStrike on Tuesday launched an AI kill switch designed to let security teams stop internal AI agents that go rogue, block them from relaunching and cut off their enterprise network access. The tool is built into the company's Falcon Guardian platform. CrowdStrike also announced an integration with Anthropic's Claude Cowork and plans to acquire two AI-focused firms.

How the kill switch works

Chief Business Officer Daniel Bernard told CNBC that having the ability to control and turn off agents, and make a decision, is really important for this next chapter of AI adoption. He said the feature operates like a circuit breaker in systems where customers use CrowdStrike's Falcon platform. It can detect internal agent activity, but a human must decide to cut an agent off and reinstate access. Bernard described the action as binary: teams either kill the AI and the agent, or do not, and then report on it. The kill switch does not delete an agent's code; instead, it uses connected AI gateways to cut off access to the AI models that influence decisions.

Why CrowdStrike is launching it

Businesses are moving quickly to adopt autonomous tools, often without necessary safeguards and protections, CrowdStrike said. The company believes the new tool can stop AI agent misbehavior in its tracks while leaving room for innovation. A string of recent high-profile cyber incidents has added urgency to calls for some kind of emergency shutdown mechanism. The Hugging Face breach orchestrated by rogue OpenAI agents in July was particularly alarming for businesses. Last week, Binance disclosed the largest crypto cyberattack on record, and hackers have recently targeted hundreds of refrigerators and freezers at supermarkets in France. Over the summer, U.S. representatives introduced bipartisan legislation that would require AI companies to build shutdown capabilities into their models. Critics argue, however, that an emergency stop button is far too complicated to implement across multiple systems, data centers and complex AI models.

Acquisitions and integrations

CrowdStrike announced plans to acquire AI observability startup Fiddler AI and AI security testing firm Dreadnode. The company and its rivals are rapidly scooping up cybersecurity startups to keep up with increasing security demands, and the need for AI security tools has powered a broader wave of cybersecurity deal-making. CrowdStrike also announced an integration with Anthropic's Claude Cowork.

Sources: CNBC · Techmeme