Updated
Updated · TechCrunch · Nov 6
Censys Finds 800 Artex AI Hacking Tools Exposed Online After July Release
Updated
Updated · TechCrunch · Nov 6

Censys Finds 800 Artex AI Hacking Tools Exposed Online After July Release

3 articles · Updated · TechCrunch · Nov 6

Summary

  • More than 800 Artex deployments were reachable on the public internet without passwords, giving outsiders potential access to autonomous hacking systems and their stored attack data.
  • Censys said the exposures appear to stem from misconfiguration, because Artex does not need to be internet-facing; some instances may be used legitimately, but any exposed copy can be misused.
  • Artex has already been tied to attacks on South Korean financial firms and the ConnectWise breach, where Microsoft said hackers used it to move through cloud environments and steal data quickly.
  • The findings highlight how simple weaknesses—open directories, weak passwords and missing multi-factor authentication—can expose both victims and the attackers' own infrastructure.
  • Artex has said it will go closed-source, but dozens of code copies remain on sharing sites, leaving the tool available for continued use in active attacks.

Insights

How does a powerful AI hacking tool end up exposing the very cybercriminals using it to attack global networks?
If autonomous AI can breach cloud systems in under three hours, are traditional security measures completely obsolete?