Hackers Use AI Agents to Do Hacking for Them

 | 

The trend was uncovered in an investigation by cybersecurity firm NorForge, findings from which were published Monday (Oct. 19).

The company’s investigation into publicly exposed artificial intelligence (AI)-driven penetration-testing control panels has found how they are widely used in malicious cyber operations. Among these panels is Cairn, the report said.

“In this setting, Cairn is an AI-powered hacking system,” NorForge said. “Give it a target and an objective, and it can coordinate AI agents to investigate the target, search for security weaknesses, and attempt to break in.”

According to the report, more than 20 exposed Cairn panels were identified as likely connected to malicious activity, offering an “unusually detailed view” into how bad actors choose targets, direct autonomous agents, exploit weaknesses and gain persistent access to compromised systems.

“It’s a new world now,” Snorre Fagerland, principal threat researcher at the cybersecurity company NorForge, told Reuters in a report on his company’s findings.

While NorForge did not identify victims by name, Reuters said it was able to determine their identities through an independent review of the hacker’s activity on Cairn. The targeted websites included farm equipment firm Deere, home appliance company SharkNinja and engine manufacturer Cummins.

Deere told Reuters it had taken “necessary remediation actions” after the incident and had “no indication that sensitive information was accessed.”

SharkNinja said the hack had involved a standalone site featuring only product support information and that the threat was contained.

Cummins was “investigating the report of a potential cyber-attack on a website operated by one of our third-party vendors,” the Reuters report added.

The findings come at a moment when AI is turning cybersecurity from an arms race into “a war of attrition, where the winner may be whoever can afford to keep fighting the longest,” as PYMNTS wrote last week.

This came after Sarah Breeden, deputy governor of the Bank of England, warned that advances in AI meant the technology now presents a “systemic risk” to the global financial system. One day earlier Bank of America CEO Brian Moynihan said his company had deployed 1,000 contractors and 3,000 employees to repair software flaws found by Anthropic’s Mythos and other frontier AI models.

“The emerging imbalance is straightforward,” PYMNTS wrote. “An attacker can deploy inexpensive AI agents to probe software, identify weaknesses and test potential exploits around the clock. A company defending that software must determine which findings are credible, investigate whether systems are exposed, prioritize remediation, deploy patches and verify that nothing breaks in the process.”