By Hamza Faisal
A Chinese-speaking hacker spent the past couple of weeks breaking into the websites of more than 100 companies — including farm-equipment maker Deere, appliance brand SharkNinja and engine manufacturer Cummins — and barely lifted a finger. According to an exclusive investigation by Reuters, the cybercriminal outsourced the hacking spree almost entirely to artificial intelligence.
Instead of typing out attacks by hand, the hacker tasked AI agents with spending hours autonomously probing and breaking into targeted sites — all for a few thousand dollars in total cost, according to Snorre Fagerland, principal threat researcher at the cybersecurity company NorForge.
Caught by watching the tools
Fagerland was not looking for this particular hacker at all. He was scanning the web for inadvertently exposed instances of Cairn, an open-source AI problem-solving platform that has become popular with hackers. Beginning around October 2, he found about two dozen instances of Cairn being used for malicious activity — which allowed him to quietly peer over the cybercriminals' shoulders as they set their AI tools loose on the world.
Fagerland's report did not name the victims, but Reuters independently reviewed the hacker's Cairn activity and was able to establish the victims' identities. The incident also highlights a new weak point in the AI supply chain itself: misconfigured, internet-exposed instances of automation platforms are now a treasure trove for anyone — attackers and researchers alike — who knows how to scan for them.
Hours of autonomous break-ins
What makes the case stand out is how little human involvement was needed. The agents were given targets and left to work for hours — probing defenses, finding weaknesses and pushing through — completing complex intrusions from start to finish without human intervention. The result, Fagerland says, is a dramatic boost to the power of a single malicious actor.
"It's a new world now"
The findings land amid rising anxiety over the power of autonomous AI tools, and they offer real-life evidence that commercially available AI agents can now pull off sophisticated intrusions on their own. Reuters said it could not ascertain the degree to which the breaches resulted in unauthorized extraction of data — but the access itself is the alarm bell.
Security teams say the lesson is blunt: exposure is no longer just a matter of servers and databases. AI agent platforms, deployed carelessly, hand intruders both the tools and the window to watch them work. One person, a few thousand dollars, and an AI workforce that never sleeps — for defenders, the math just got terrifying.
