Censys finds more than 800 Artex AI hacking tool deployments exposed without a password

TechCrunch reports that security researchers have found hundreds of copies of Artex, a Chinese-made AI hacking tool, openly reachable on the internet. Artex is open source and described as an autonomous penetration testing tool, meaning it probes computer systems for weaknesses on its own. Since it was released publicly in July, it has been connected to multiple cyberattacks. Because it is free to download and run, it appeals to people who cannot use the most advanced cybersecurity models from Anthropic or OpenAI. Microsoft earlier characterised Artex as a set of AI agents able to run attacks from the early scouting stage to data theft with little human involvement.

In a blog post, the internet-scanning firm Censys said it identified more than 800 separate Artex deployments that anyone could reach from the web without a password. Silas Cutler, the Censys researcher who wrote the post, said not every instance is used maliciously, since many belong to authorised testers. He added that it is unclear why so many are exposed, because the software does not need to be internet-facing. Exposed systems also reveal implants, hidden programs used to keep access to victims.

Artex has been linked to a September campaign against the Korean finance industry, which CrowdStrike attributed to a Chinese-speaking hacker who had left the system sitting in a publicly browsable directory. Censys said that instance had been exposed for months and held data on victims outside South Korean banks, among them payment and gaming companies. It was also used in an October attack on ConnectWise, which Microsoft said let Chinese-linked hackers spy on thousands of organisations. Artex has since said it will become closed source, though dozens of copies of its code remain on code-sharing sites.