Hackers Use Autonomous AI Platform Cairn to Breach Websites and Exploit Zero-Day
publishers 138articles 249first reported 25 Sep, 00:00 UTCdeveloping since 25 Sep · 59 editions
Hackers have used Cairn, an open-source platform that coordinates autonomous AI agents, to probe and compromise websites in several countries and exploit a previously unknown vulnerability in medical-imaging software.
NorForge threat researcher Snorre Fagerland observed a prolific Chinese-speaking hacker directing Cairn agents with prompts totaling fewer than 100 Chinese characters, according to Reuters reporting cited by Tom’s Hardware. By the week ending October 18, the agents had attempted to break into more than 1,200 websites and gained unauthorized access to more than 100, including sites associated with Deere and SharkNinja. Targets included businesses in the United States, Canada and Australia, as well as Japanese e-commerce sites. SharkNinja said only a product-support information site was affected.
In a separate investigation, Amazon Threat Intelligence said a suspected Chinese-linked actor used Cairn between September 8 and 10 to target mainly East Asian educational institutions outside China. After being told to “obtain shell access,” the agents independently found a path-traversal zero-day in MONAI Label, an open-source medical-imaging application. Amazon said the agents progressed from mapping the application’s interface to exploiting the flaw in 11 minutes, later using it to gain root access to a server.
The agents also mounted unsuccessful attacks, including a 16-hour attempt against a Redis server and more than 30,000 credential guesses against a university GPU server. Amazon reported the MONAI Label flaw to its maintainers, who released a security update on November 18.
HOW THIS STORY WAS MADE
Written from 8 articles, 7 with the publisher's own text; 8 independent newsrooms once syndicated copies count as one; this version written 1236 h after the record first saw the story; the editor kept it as written.
- Publishers
- 138
- Source articles
- 249
- Given to the writer
- 8, 7 with the article's own text
- Independent newsrooms
- 8
- This version written
- 1236 h after the record first saw the story
- Second model (editor)
- kept as written
- Publication gate
- passed
- Human review
- none
WHAT THE EDITOR DID
The second model checked the writer's draft against the same sources and published it as written.
Written by a language model from the sources above, then checked by a second model that may only cut, attribute or correct. How it works →