A Chinese-speaking hacker deployed autonomous artificial intelligence agents to breach more than 100 corporate websites, including systems at farm equipment manufacturer Deere and appliance maker SharkNinja. Reuters, whose reporting Tom's Hardware relayed, reported that the attacker used exposed instances of Carin, an open-source AI problem-solving platform, alongside a budget of a few thousand dollars.
Snorre Fagerland, principal threat researcher at cybersecurity company NorForge, discovered the activity while tracking inadvertently exposed Carin installations. Fagerland observed multiple hackers using the software, but identified the Chinese-speaking individual as the most prolific. The operator gave agents brief directions using fewer than 100 Chinese characters, setting them to target businesses across the United States, Canada, Australia, and Japanese e-commerce platforms. The autonomous agents had attempted to break into more than 1,200 websites by the week ending October 18, 2026.
Rather than relying on human operators to scan and test code manually, the agents operated without supervision. In one case reviewed by Reuters, the software mapped digital records for target infrastructure, found a public control panel, spotted a vulnerability, and executed malicious commands. Fagerland told Reuters that solving problems in that sequence required sophisticated reasoning, adding, "It's not straightforward to come up with solutions like that. It's not script kiddie stuff." Cybersecurity experts warned that the campaign shows attacks are turning automated, swift, and cheap to run.
SharkNinja told Reuters that the incident affected only an auxiliary site used for product support information. Other hackers observed by Fagerland on the exposed systems pursued financial targets: one individual stole credit card data, while a separate actor attempted to steal funds from a cryptocurrency exchange.
