Security researchers have discovered that hundreds of copies of AI hacking tools manufactured in China have been publicly exposed on the internet, accessible to anyone, which may enable outsiders to launch cyberattacks against unprepared targets.
This hacking software, named Artex, is an open-source automated penetration testing tool that can automatically detect vulnerabilities in computer systems. Since its public release in July of this year, it has been associated with several cyberattacks. As open-source software, it is available for anyone to download for free and run on their own servers, making it a favored tool for those who cannot use the more advanced network security models Anthropic or OpenAI to identify security flaws. Microsoft has previously described Artex as a set of AI proxy systems that can launch cyberattacks from reconnaissance to data leakage, with almost no need for human intervention.
A cybersecurity company, Censys, stated in a blog post that when scanning for systems exposed to vulnerabilities or with existing vulnerabilities on the internet, they discovered more than 800 different Artex instances that were accessible via web pages and did not require a password for login.
The chief security researcher Censys who wrote this blog, Silas Cutler, stated that not every Artex instance is used maliciously; many are for authorized testers. However, it is still unclear why hundreds of Artex systems are exposed to the internet, as the software does not require them to be accessible from the internet.
These findings highlight that even powerful hacking tools can become publicly accessible to anyone due to improper configuration, thereby increasing the risk of being discovered and misused by others. Publicly accessible systems can also expose attack tools, such as implantation programs (hidden malware) used to infiltrate and maintain continuous access to the victim's network.
Security companies have linked several large-scale cyberattacks to the use of Artex, including an attack on South Korea's financial industry in September. Cybersecurity giant CrowdStrike attributed these attacks to a Chinese-speaking hacker, stating that the hacker used an instance of Artex during the attacks. Part of the reason is that this hacker left an "open directory" that hosted the system. Censys indicated that their data shows that the involved Artex instance has been publicly exposed for several months and seems to contain more attack data, affecting a range "far beyond South Korean banks," also including payment and gaming companies.
Artex was also used in an attack targeting IT in early October of this year, with the support of giant ConnectWise. This allowed hackers associated with China to monitor thousands of organizations whose data and files were hosted on Microsoft's cloud. Microsoft was aware of this attack and stated in a blog post that the hackers used Artex to invade customers' cloud environments and steal data at such a rapid pace that defenders "had almost no time to detect and respond."
Many attacks related to Artex have exploited simple vulnerabilities, such as guessing weak passwords or known passwords, as well as the lack of multi-factor authentication.
These findings indicate that simple vulnerabilities, including configuration errors, may simultaneously expose both the attacker and their target.
Artex has subsequently announced that it will become closed-source, which effectively limits further releases to those who are not paying customers. However, there are still dozens of copies of the Artex code on code-sharing websites that can still be used for ongoing attacks.
