Hackers using an artificial intelligence tool gained high-level access to some organizations’ computer systems in less than three hours, the FBI warned Friday. The clock started when the hackers first picked their target. Even when the activity triggered an alert, security teams had little time to investigate before the attackers were inside.
The tool, called Artex, became publicly available in July as a way to test computer networks for weaknesses. Within months, the FBI said, hackers had used it in attacks on South Korean financial institutions and customers of ConnectWise, an American IT management company. Government agencies, defense contractors, telecommunications providers and law firms were also among the targets.
On Monday, Google warned that hackers were also using AI to hunt for targets without waiting for human direction. Its researchers documented an operation that monitored software updates around the clock. When a company released a security fix, the AI could identify the flaw the update had closed and develop an attack against systems still running the older software. For organizations that took days to install an update, part of that time had once been spent waiting for a hacker to figure out what the patch revealed. The software could now do that work on its own.
In the ConnectWise case, explained TechCrunch, the opening came through a backup application that customers had already authorized to access their Microsoft 365 accounts. Microsoft attributed the October attack to a group it tracks as Storm-0940, which it described as aligned with the Chinese government. The compromised application gave the group a route into customer systems through software the customers already trusted.
The FBI put the reach of the compromised application at more than 2,250 customer environments. Microsoft confirmed that the hackers signed in and carried out further activity in multiple environments, though the number of customers whose data was stolen remains unclear. A ConnectWise spokesperson told TechCrunch the company could not independently confirm how many were affected.In South Korea, financially motivated hackers stole personal information belonging to about 67,000 people in attacks on financial institutions, according to the FBI. Some of the affected systems lacked multifactor authentication, a safeguard that requires users to verify their identity a second way when signing in.
At Barracuda, a cybersecurity company that disclosed a separate Artex-linked breach last week, the opening came from a former vendor. Barracuda had ended the business relationship but left a credential active. An attacker used it to reach systems where Barracuda kept credentials for monitoring customers’ networks. The company’s records could not establish which credentials the attacker had taken, so Barracuda replaced those within reach and asked affected customers to replace theirs.
The attacker moved across five platforms in less than two days, according to Arve Kjoelen, Barracuda’s chief information security officer. The most useful tip Barracuda received from another company arrived two days after the attack began. “Defenders who rely on spotting malicious activity and then responding to it will often arrive after the attacker has finished.”
[READ MORE: Why Gerrymandering Doesn’t Work]

