FBI Warns Artex AI Tool Breaches Networks in Under Three Hours

The Federal Bureau of Investigation warns that hackers are using autonomous artificial intelligence tools to launch cyberattacks across the United States and globally, moving from initial target selection to active network compromise in under three hours. These campaigns have affected several industries, including financial services, government, technology, defense, telecommunications, internet companies, and law firms.
The advisory highlights an open-source autonomous penetration testing platform named Artex, which probes computer network defenses. Operators can connect Artex to hosted frontier large language models from providers like Anthropic and OpenAI, alongside open-source and non-U.S. models. The software rapidly scans victim networks for weaknesses, exploiting simple security flaws such as weak protections and exposed or stolen passwords to secure an initial foothold. From there, the tool enables attackers to navigate through the network searching for databases, mailboxes, and sensitive information. While these automated intrusions generate heavy traffic and can trigger defensive alerts, they operate at speeds that human defenders struggle to match.
The FBI notes that Artex has been deployed in multiple cyberattacks since its public debut in July. In South Korea, hackers used the software to breach financial institutions, exploiting weak security defenses and a lack of multi-factor authentication to compromise the personal data of at least 68,000 people. The FBI also confirmed findings from Microsoft that hackers used Artex to break into IT management giant ConnectWise, gaining access to the Microsoft 365 cloud environments of at least 2,250 customers. Microsoft reported that these hackers were likely Chinese actors who used the tool to steal data from affected organizations.
A spokesperson for ConnectWise told TechCrunch that the company does not have the visibility to confirm the number of affected environments cited by the FBI. Cybersecurity company Barracuda also disclosed a separate breach involving Artex, where financially motivated criminals located an exposed credential belonging to a third-party service that Barracuda had not revoked after ending its work with the vendor. Because its internal logs were insufficiently detailed to rule out unauthorized access, Barracuda chose to assume the worst and reset all customer credentials.
Arve Kjoelen, chief information security officer at Barracuda, wrote in a statement: "Defenders who rely on spotting malicious activity and then responding to it will often arrive after the attacker has finished". A Barracuda spokesperson told TechCrunch that the company directly notified and advised a minority of customers using its extended detection and response services to revoke and recreate integration credentials, though it did not specify the exact number.
Following these incidents, the developer converted Artex into a closed-source project to restrict further misuse, but the FBI reports that the prior open-source version continues to circulate and remains widely accessible. Security researchers also discovered hundreds of publicly exposed Artex instances online that had been placed on the internet without passwords due to misconfigurations. Among those exposed deployments was the specific Artex system used to target South Korea.
Summary
- Automated AI penetration tools can bypass traditional human response times by completing network compromises in under three hours.
- Open-source versions of powerful hacking software continue to circulate online despite developers moving the project to a closed-source model.
- Unsecured deployments left without passwords have accidentally exposed advanced cyberattack capabilities to the wider internet.