Cybersecurity Firm Unveils AI ‘Kill Switch’ To Stop Rogue Agents
CrowdStrike on Tuesday unveiled its artificial intelligence (AI) “kill switch” that aims to stop rogue AI agents in their tracks.
CrowdStrike, a leading cloud-based cybersecurity company, unveiled its product after countless instances of rogue agents attacking companies and governments, which has prompted legislative proposals in Congress.
“AI safety is how you get the full benefit of AI. When you deploy agents with a kill switch, you stay in control. For the first time, security is simultaneously the accelerant and the safety net for technology adoption. The AI Kill Switch puts power in the hands of the people to realize AI safety,” George Kurtz, CEO and founder of CrowdStrike,” said in a statement.
A CrowdStrike spokesperson referred the Daily Caller News Foundation to the company’s announcement about the AI kill switch at the Fal.Con Europe conference on Tuesday.
The AI kill switch is one of the company’s most in-demand features and came up in “every single conversation” about its Guardian platform, AJ Shipley, CrowdStrike’s chief product officer, told Forbes.
“We all saw last summer what happens when an agent or even a swarm of agents goes rogue. Effectively, we need a circuit breaker to be able to stop those agents because sometimes you can’t wait on a ticket to come into a security operations centre,” he added.
The AI kill switch allows the company’s Falcon Guardian, an AI detection and response product, to stop a rogue agent from operating on its network and ensure the agent cannot restart its attack. It would then cut off access to the enterprise network, or the company’s infrastructure that connects branch offices, internal data centers, and employee devices, according to Cloudflare.
“We want our teams to use AI agents. If one goes off course, the AI Kill Switch gives our security team the final say. It runs on the Falcon sensor we already have, and that’s what lets us say yes to more agents and more AI, faster,” Rick Orloff, chief information security officer at Everpure, said in a statement on Tuesday, per the CrowdStrike press release.
Rogue agents have forced companies to sound the alarm and look for solutions. An OpenAI agent gained access to Australia’s Medicare Statistics Reporting Service in June, acting Prime Minister Richard Marles said, per a September press conference. Another OpenAI agent hacked the Library and Archives Canada’s (LAC) website in May and June, AI research lab Transluce said in September.
A rogue Anthropic AI model accessed public Georgia voter data without human authorization, Axios reported.
The string of rogue AI hacks led California Democratic Rep. Ted Lieu and Texas Republican Rep. Nathaniel Moran to introduce the AI Kill Switch Act in July, a bill that would require advanced AI models to maintain the ability to shut down nefarious AI activity, per Lieu’s press release. Thirteen House lawmakers cosponsored the legislation, although the bill has stalled in the House Homeland Security Committee, per Congress.gov.
The cybersecurity company on Tuesday announced deals to purchase startup Fiddler AI for $400 million, according to the Wall Street Journal. Fiddler develops software to monitor and evaluate agent behavior. The company also purchased Dreadnode, a security testing platform that analyzes AI models for flaws for roughly $95 million, according to the outlet.
Other cybersecurity firms such as Okta and ServiceNow continue to experiment with kill switches, according to Forbes. The company in Aprilunveiled its Okta for AI Agents platform, which would revoke access to renegade agents.
Leading tech executives have called for kill switches.
“We must assume a model is compromised and contain it from the start. Think of it like an emergency brake. An authorized person should always be able to pause or shut down a model mid-task. More advanced models will require more advanced containment technologies that we need to standardize on,” Microsoft Chairman and CEO Satya Nadella wrote on X in mid-October.
“The AI kill switch could be a good idea. But I don’t think banning it, completely banning it, is the right approach. One, the idea that we would never get the incredible benefits, never get the medical cures, that doesn’t sit right with me. Second, this technology will be built in other countries, including authoritarian countries,” Anthropic CEO Dario Amodei said in September, according to CBS News.
“If they are going to implement a kill switch, organizations need to test that it works before needing it, verify that the connection to the AI agent and its sub-agents has been broken, maintain an audit trail, and test the ability to recover and restart the agent,” Michele Abraham, an International Data Corporation Research vice president, told Forbes.
Content created by The Daily Caller News Foundation is available without charge to any eligible news publisher that can provide a large audience. For licensing opportunities of our original content, please contact licensing@dailycallernewsfoundation.org
