AI cyberattack: a hacker used AI agents and $5,000 to attack more than 100 enterprises

According to an investigation by NorForge researcher Snorre Fagerland, a Chinese-speaking hacker used the open-source multi-agent platform Cairn to direct AI agents in attacks on more than 100 enterprises in the United States, Canada, Australia and Japan. The attacks began around September 25; the agents attempted to breach more than 1,200 websites, and in successful cases vulnerability search and analysis typically lasted about 23 hours; the entire operation is estimated to have consumed about $5,000 in AI tokens, roughly $50 per successful intrusion, and is said to have combined the DeepSeek model with Anthropic's Claude Code. Reuters, through independent analysis, confirmed that affected enterprises include Deere, SharkNinja and Cummins; the three companies said respectively that they had taken remedial measures, that the threat was contained, or that they were investigating.