One hacker used AI agents to break into more than 100 organisations' websites for about $5,000, Reuters reports

A Chinese-speaking hacker used commercially available AI agents against more than 1,200 websites and got into more than 100 organisations, Reuters reported on 19 October, according to Startup Fortune and SOFX. Reuters itself would not open for us, so the figures here are Reuters', as those two outlets carry them. Both put the attacker's spend at about $5,000 in AI tokens. SOFX says that works out at roughly $50 and about 23 hours of automated probing for each successful break-in.

The attacker was found by Snorre Fagerland, a threat researcher at the security firm NorForge, while he was searching for exposed deployments of Cairn, an open-source platform that coordinates AI agents. SOFX says Fagerland's report did not name victims, and that Reuters identified Deere, SharkNinja and Cummins on its own.

Deere said it took “necessary remediation actions” and had seen no sign that sensitive information was accessed. SharkNinja said a standalone product-support site was hit and the threat was contained. Cummins said it was looking into a website run by a third-party vendor. Reuters could not determine how much data was taken, Startup Fortune reports.

The $5,000 is an estimate. The outlets disagree on which AI models the attacker used, so the Switchboard names none.

If the estimate holds, it puts a price on a break-in by AI agents: tens of dollars of tokens for each ordinary company website, run from short instructions by one person.

• Reported The hacker targeted more than 1,200 sites, breached more than 100 organisations and spent about $5,000, Reuters reported, according to Startup Fortune. Startup Fortune
• Reported NorForge's Snorre Fagerland found the hacker through exposed Cairn instances, at about $50 and 23 hours per breach, Reuters reported, according to SOFX. SOFX
• Claimed Deere says it took “necessary remediation actions” and saw no sign sensitive information was accessed. Startup Fortune