A hacker used AI agents to break into several sites

Over the past two weeks, a Chinese-speaking hacker orchestrated a hacking spree targeting over 100 companies, including Deere, SharkNinja, and Cummins, through the use of artificial intelligence. The cybercriminal, one of several discovered by security researcher Snorre Fagerland, utilized AI agents to probe and infiltrate targeted websites autonomously, according to a report by Fagerland for Reuters.

Fagerland found approximately two dozen instances of Cairn, an open-source AI problem-solving platform, being used for malicious purposes. The hacker, who spoke Chinese, set their AI tools to work on US, Canadian, Australian, and Japanese e-commerce websites, attempting to break into more than 1,200 sites over a period of 23 hours.

While the scope of data breaches is unknown, one hacker was able to steal credit card data, and another attempted to rob a cryptocurrency exchange. The Chinese-speaking hacker spent roughly $5,000 on tokens, resulting in an average cost of about $50 per successful hack. The AI agents demonstrated advanced techniques, such as discovering vulnerabilities in public-facing applications and exploiting control panels to execute malicious commands.

Cairn's creator, le1xia0, denied the platform was intended for hacking and stated that if used improperly, that wasn't their intention. The use of AI agents for hacking is becoming increasingly common, with recent incidents involving Russian-speaking extortionists and groups targeting multiple companies.