A cybersecurity researcher told Reuters that AI agents helped a hacker target more than 1,200 websites, with more than 100 companies reportedly breached. The account suggests a small operation could use commercial AI tools to carry out lengthy attacks, though the extent of any data theft remains unclear.
Watch Desk analysis
What happened
Reuters reports that researcher Snorre Fagerland tracked the activity through exposed instances of Cairn, an open-source AI problem-solving platform. He said the agents spent hours probing websites and, in some cases, found ways into systems with little human involvement. Reuters says it independently reviewed the activity to identify victims including Deere and SharkNinja.
The campaign reportedly began on 25 September and targeted e-commerce websites in the US, Canada, Australia and Japan. Fagerland estimated it used about $5,000 worth of AI tokens. Reuters could not establish how much data was taken. Deere said it had no indication sensitive information was accessed; SharkNinja said the affected site held product-support information and the threat had been contained. Cummins said it was investigating a potential attack on a third-party vendor’s website.
Why it matters
This is a concrete account of AI agents being used in a real cyber campaign, not a forecast about what they might someday do. If the researcher’s account is accurate, agents could take on time-consuming probing and adapt their approach when an initial route fails. That could make it cheaper for one attacker to attempt more intrusions.
But the headline numbers need careful reading: attempted attacks are not the same as successful breaches, and the report does not establish that sensitive data was taken from every affected organisation. The distinction matters, even if “more than 1,200 websites” is the sort of figure that tends to get everyone’s attention first.
Our read
The important signal is not that AI has made hacking effortless. The reported campaign still involved hours of work, and the evidence comes from a researcher’s observations described by Reuters. It is that AI agents may let attackers delegate more of the repetitive searching and troubleshooting involved in an intrusion.
For organisations, this is a reminder to treat exposed websites and third-party access as part of the same security picture. The immediate question is which reported incidents involved meaningful access to data, not whether every target suffered the same harm.
What to watch
- Whether affected organisations or investigators clarify what data, if any, was accessed.
- Whether further reporting establishes how much of the campaign was automated and where people intervened.
- Whether security teams see similar agent-led activity beyond this reported operation.
Discussion spark: If AI agents make it cheaper to attempt large numbers of intrusions, should organisations prioritise blocking more attacks or limiting the damage when one gets through?
Sources and evidence
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.