FBI Warns That AI Hacking Tools Such as Artex Can Obtain Target Network Privileges in Under 3 Hours
The U.S. Federal Bureau of Investigation (FBI) warned in a cybersecurity flash bulletin that hackers are relying on open-source autonomous AI hacking tools such as Artex to launch attacks, with some attacks going from locking onto a system to completing the intrusion in under three hours. Artex can connect to frontier large models such as Anthropic and OpenAI, first scanning the target network to find weaknesses, then using weak protections, exposed or stolen passwords to gain a foothold and spread laterally to databases and email. The bulletin said that multiple South Korean financial institutions were breached due to the lack of basic protections such as multi-factor authentication, and the personal data of at least 68,000 people was stolen; Microsoft previously disclosed that attackers used the tool to breach ConnectWise and thereby enter the Microsoft 365 cloud environments of at least 2,250 customers.
Why read this
AI score 69
The FBI bulletin disclosed the attack paths and scope of impact of AI hacking tools, which can be used to understand the real-world risks of autonomous agents in attack and defense scenarios.
Topics
Body · AI translation
FBI: AI Hacking Tools Like Artex Can Obtain Target Network Privileges in Under 3 Hours
Phoenix Tech Download the App for Exclusive Early Access
November 17, 2026, 07:00:09 From Beijing
Phoenix Tech News, November 17 (Beijing Time) — According to TechCrunch, the U.S. Federal Bureau of Investigation (FBI) warned in a recent cybersecurity flash bulletin that hackers are relying on open-source autonomous AI hacking tools such as Artex to launch attacks, affecting customers of U.S. IT management company ConnectWise and South Korea's financial sector.
Artex is a type of open-source AI penetration testing platform that can connect to frontier large models such as Anthropic and OpenAI, as well as open-source models and non-U.S. models. The FBI said these tools first rapidly scan target networks for weaknesses, then use weak protections, exposed or stolen passwords to gain a foothold, and subsequently spread laterally to databases and email inboxes. Some Artex-related attacks took "less than three hours" from locking onto a system to completing the intrusion.
The bulletin disclosed that since Artex was first publicly revealed in July, it has been used in multiple attacks. Several South Korean financial institutions were breached due to a lack of basic protections such as multi-factor authentication, and the personal data of at least 68,000 people was stolen; Microsoft previously disclosed that attackers used the tool to breach ConnectWise and thereby gain access to the Microsoft 365 cloud environments of at least 2,250 customers. Security company Barracuda also disclosed that it suffered an Artex-related attack and, due to insufficient logs, "proceeded on a worst-case basis" and notified a small number of potentially affected XDR customers to revoke and recreate relevant credentials.
The FBI also noted that the Artex developer has since made the software closed-source following the attacks, but old open-source versions are still circulating and are "widely available." Last week, researchers also found hundreds of Artex instances exposed on the internet without passwords — including the one used to attack South Korea. Barracuda Chief Information Security Officer Arve Kjoelen said: "Defenders who rely on discovering malicious activity and then responding often arrive only after the attackers have packed up and left." (Editor/Qiongqi)
For more first-hand news, please download the Phoenix News app and subscribe to Phoenix Tech. For in-depth reports, search "Phoenix Tech" on WeChat.
AI Trends