Donald Trump’s FBI has issued an emergency warning about a free hacking tool that can give criminals control of an entire company network in less than three hours.
The bureau identified the tool as ‘Artex’ and said hackers have already deployed it against banks, government agencies, energy companies and hospitals.
Artex is described as “an open-source, large language model (LLM)-driven agentic platform” that threat actors have weaponized “to automate multiple phases of cyber intrusions at machine speed,” according to the FBI.
The alert said the technology “significantly reduces the time required to chain attack techniques.” It allows people with limited technical expertise to conduct sophisticated attacks against multiple targets simultaneously.
The FBI warning follows comments from Federal Reserve Governor Michael Barr, who told a conference in Italy earlier this week that AI agents could drain a bank before regulators had time to respond.
Barr said Monday that AI could move customers’ money automatically and at extraordinary speed, creating what he called a potential “stampede.” A surge in withdrawals could overwhelm a bank’s ability to manage its cash reserves “in minutes,” he warned, leaving supervisors with no time to intervene.
Major banks are already detecting AI agents attempting to operate inside retail customer accounts, Barr said, while institutions have no reliable way to determine who authorized the activity. An FBI spokesman said Tuesday, however, that the Artex alert was unrelated to the recent series of widespread cyber outages.
In recent days, US cyberinfrastructure has been hit by email disruptions, a telecommunications failure that grounded flights nationwide and outages affecting online banking services.
The FBI identified the tool as ‘Artex’ and said hackers have already deployed it against banks, government agencies, energy companies and hospitals
The alert said Artex “significantly reduces the time required to chain attack techniques,” enabling users with limited technical expertise to run complex operations against multiple targets
US cyberinfrastructure has faced email disruptions, a telecommunications failure that grounded flights nationwide and outages affecting online banking services
The White House said the outages were not caused by a cyberattack, attributing them instead to updates involving “old computer systems.”
More than 450 million Microsoft 365 users lost access to email last week. On Friday, the Federal Aviation Administration grounded flights across the country — the second nationwide stoppage since 9/11.
MORE STORY: Diagnosed With Parkinson’s at 36, I Found a Radical Fix
By Monday, Chase customers were unable to access banking services and United could not contact its pilots. The following day, the National Security Agency’s internal network also went down.
Amazon said Friday that it had detected hackers using Artex to steal credentials for AI models before using the compromised access to search for additional targets.
Google’s Threat Intelligence Group, meanwhile, published a report Monday warning of “a paradigm shift towards intrusions occurring at machine speed.”
The group said it had observed hackers using AI to identify software vulnerabilities, create attacks and deploy them automatically, without human involvement — something it said had happened for the first time.
Google said AI-powered attack tools were inexpensive, straightforward to operate and capable of being deployed at enormous scale.
“Imagine a hacker that doesn’t eat, sleep, or get tired, and never stops searching for a way in,” Google said in a statement.
Federal Reserve Governor Michael Barr said Monday that AI could move customers’ money automatically and at high speed, describing it as a possible “stampede”
Hackers have compromised some of the world’s most visited websites in a sweeping cyberattack targeting users’ money and credit card details
“According to the latest AI Threat Tracker report, bad actors are now using autonomous LLM [Large Language Model] frameworks to continuously scan software updates, determine which vulnerability was patched and build an exploit for it — all without human involvement.”
The report found that attacks exploiting newly discovered, previously unknown software flaws averaged eight per month in 2025 before rising sharply to 22 in August.
Suspected North Korean hackers reportedly posed as recruiters and used AI-generated profile images to target employees at cryptocurrency companies in North America and Europe.
Chinese state-linked spies used AI during the summer to investigate targets, identify weaknesses and develop attacks. A separate tool known as CAIRN was used against organizations in the US aerospace industry.
In another incident, investigators found Russian-language notes embedded in a system’s files. Google said the operator appeared to be targeting websites run by French media, entertainment and education organizations.