Washington, Nov 14 — The FBI has warned that cybercriminals are using an artificial intelligence-assisted security testing tool to accelerate attacks on organisations around the world, with some intrusions reaching privileged access in less than three hours.
In an alert issued on Friday (local time), the bureau said malicious actors had used the open-source framework, known as ARTEX, since at least August to target organisations across sectors including finance, government, technology, defence, energy and health care.
The FBI said the tool can automate several stages of an intrusion, from identifying potential targets to preparing or carrying out data theft. That speed can leave defenders less time to detect and respond to an attack, it said.
The alert said some intrusions progressed from initial targeting to privileged access in under three hours.
According to the alert, the framework was originally intended for authorised security testing. Its developer removed the original online repository in October following reports of malicious use, but copies and variants remained available, the FBI said.
The bureau said the tool could lower the technical barriers for conducting complex operations and allow attackers to pursue multiple targets at the same time. Organisations' internet-facing applications, cloud infrastructure and systems connected to suppliers or partners were among the areas of concern identified in the alert.
The FBI cited financially motivated attacks on South Korean financial institutions that it said resulted in personal information being taken from approximately 67,000 people. It also described a separate supply-chain compromise affecting users of a cloud backup service, which it said enabled access to more than 2,250 downstream customer environments.
The bureau encouraged organisations to review their exposure to the internet, strengthen identity and access controls, monitor unusual activity and protect credentials. It also urged organisations that identify suspicious activity to contact the FBI or report it through the Internet Crime Complaint Center.