The FBI is sounding the alarm over ARTEX, a free AI hacking tool that lets criminals and foreign spies break into corporate networks in just a few hours.
In an urgent bulletin released November 13, the bureau said the program has been used since at least August against banks, government agencies, defense contractors, hospitals, energy firms, chipmakers, and law offices around the world. In some cases, hackers went from picking a target to holding the keys to the kingdom in under three hours.
The scariest part is that people using it barely need to know what they’re doing. Users type out what they want in everyday language, and a team of AI “agents” does the work. They scout for weak spots and crack passwords while taking out data. The feds warn the tool lets crooks with “minimal technical expertise” run several attacks at once.
ARTEX doesn’t have its own AI brain. Instead it plugs into chatbots like Claude, ChatGPT, and China’s DeepSeek. The FBI says hackers mostly pick Chinese models, especially DeepSeek, because they’re cheap and don’t seem to put up much of a fight when the job is described as “security testing.”
A Chinese developer who goes by “Autumn-27” posted the tool on GitHub in July and billed it as a legitimate way for companies to test their own defenses. He pulled it on October 8 after reports that hackers were abusing it, but by then, it was too late. Copies and a paid “ARTEX PRO” version are still circulating. Researchers counted more than 1,600 of its control panels sitting wide open on the internet in early November, and the real number of copies in use is probably much higher.
Microsoft says a China-aligned crew it calls Storm-0940 used ARTEX to fire off more than 151,000 login attempts, about 2,900 a minute, at eight organizations in October. When that mostly flopped, the hackers broke into ConnectWise’s cloud backup service and stole its access keys. Then they tried to break into more than 2,250 companies’ Microsoft 365 systems. Security firm Huntress confirmed at least 173 were compromised.
Hackers used ARTEX with AI coding agents to hit Shinhan, KB Kookmin, and Hana banks, exposing personal data on roughly 67,000 people.
South Korean cops arrested seven suspects tied to attacks on more than 90 organizations and the theft of about 8.2 million records.
CrowdStrike says a Chinese espionage group it calls Vapor Panda used ARTEX against more than 200 organizations in 14 countries and confirmed follow-up hacking at nine organizations.
Microsoft says it watched one ARTEX agent go from hunting for leaked passwords to stealing data from a cloud vault in five minutes flat.
And ARTEX isn’t the only one out there. Security firm GitGuardian has found 45 similar “sibling” tools.
The good news is that many of the crooks using it are sloppy and leave their own gear exposed.
“AI is lowering the skill floor for offensive operations,” said Matt Snyder of Aviatrix, but it isn’t making hackers any more careful.
The experts say the old-school basics still work. “Enforce MFA, monitor your attack surface, and practice good credential hygiene,” said Silas Cutler of Censys.
The FBI is urging companies to lock down forgotten systems like partner portals and old apps, and to report suspected ARTEX activity at ic3.gov.

.png)
.png)

