AdvertiseDonate
Friday, November 6, 2026
Today’s news, without the noise.
Live Herald Desk · 388 South Main Street, Suite 440, Akron, OH 44311
Tech · National

CrowdStrike Unveils Tool to Halt Uncontrolled AI Agents

The company's new feature allows security teams to terminate autonomous software and cut network access, a move the CEO describes as essential for enterprise safety.

By Sophie Costa ·

Illustration for: CrowdStrike Unveils Tool to Halt Uncontrolled AI Agents
Illustration generated for this article, not a news photograph.

Enterprise security has traditionally focused on blocking external hackers, but the rise of autonomous software is shifting attention inward. CrowdStrike announced a new feature Tuesday that allows organizations to stop artificial intelligence agents that behave unexpectedly.

The tool, part of the Falcon Guardian platform, acts as a kill switch for AI agents. It enables security teams to terminate misbehaving agents, prevent them from restarting, and sever their network connections. The technology is available immediately and is designed to give human operators direct control over AI systems within their infrastructure.

CrowdStrike CEO George Kurtz explained the utility of the feature on CNBC's Mad Money. He noted that many companies are deploying AI agents but lacked a simple way to stop them. "Everyone is building AI agents, but they never had the off button, and now they do," Kurtz said. He added that the feature is a key component in ensuring these systems operate safely.

The launch follows a wider industry debate about the pace of AI development. In September, Anthropic CEO Dario Amodei called for slowing AI progress due to safety concerns. OpenAI CEO Sam Altman supported that view. Kurtz argued at the time that slowing development would not remove existing risks. He maintained that safeguards are necessary as companies deploy autonomous systems.

Kurtz compared the new tool to airbags in automobiles. He argued that safeguards are required even when systems are designed to function correctly. "If they go outside of those guardrails at runtime, we can kill those agents," he said. He described enabling agents to work safely as a large market opportunity for CrowdStrike.

The company is seeing increased demand for its tools as businesses adopt AI. Kurtz cited a partnership with Anthropic, which integrates CrowdStrike technology into its Claude Cowork product. He said enterprise customers often requested this integration before deploying Anthropic's offerings. "Our customers went to Anthropic and said, 'We need CrowdStrike before we can deploy AI,'" Kurtz said.

CrowdStrike shares reached a record high of $278.86 on October 6. The stock has since declined by approximately 12% as tech stocks faced pressure in October and early November.

Source: CNBC

More Tech · Report a business · Got a Story?

Get Live Herald Desk in your inboxToday’s news, without the noise. Free, a few times a week, unsubscribe any time.

Get Live Herald Desk in your inboxToday’s news, without the noise. Free, a few times a week, unsubscribe any time.