AI Platform Cairn Abused to Hack Over 100 Organizations - SecurityWeek

Marvellous news from the open-source community: Cairn, an AI platform whose entire appeal was that anyone could pick it up, has been picked up by anyone, and rather more than a hundred organisations are now the poorer for it. Norforge — an AI security firm, an industry currently enjoying the sort of growth curve that makes ransomware look like a cottage trade — reports "three clusters of activity that appear to be malicious", which is the vendor-grade equivalent of a doctor telling you that you appear to have a leg. Persistent access and financial theft: the twin load-bearing pillars of every criminal business plan since someone first chiselled a coin.

Note the name. A cairn is a pile of stones left on a path so travellers don't get lost. Someone built one, open-sourced it, and the first people to arrive were carrying a bag of other people's credentials. The operation is automated, which means the marginal cost of robbing the hundred-and-first organisation is roughly the price of the electricity, and the platforms being hit number in the hundreds worldwide — a figure that sounds enormous until you remember the internet has hundreds of everything, including problems.

And so the disclosure ritual runs on schedule: a security firm names three clusters, a trade publication files the story, and the words "financial theft" go by without a single regulator clearing their throat. We have spent a fortnight admiring open-source generosity — Google's TorchTPU hand-out being the recent example — and here is the invoice arriving, itemised. Nobody has named the targets, nobody has said which country the hundreds of websites sit in. The three clusters are the only detail on offer: no techniques, no indicators, no names. Cairn's maintainers, one assumes, are having a perfectly ordinary Tuesday.