The Federal Bureau of Investigation (FBI) last week issued an advisory about ARTEX, an AI system it says “malicious cyber actors” have used to target numerous organizations beginning in August 2026.
The advisory was published on Friday, as a string of high-profile outages began to impact millions of Americans. However, despite the timing of the FBI’s advisory coinciding with these disruptions, officials have not linked any of the outages to ARTEX or indicated that any of the disruptions were related to cyberattacks or artificial intelligence.
The disruptions began Thursday with a widespread Microsoft 365 outage and continued Friday, when there was a rare, hourslong nationwide Federal Aviation Administration (FAA) ground stop. It continued into this week, with problems affecting major U.S. banks on Monday. Also on Monday, hackers exploited a vulnerability in website-optimization platform Optimizely to inject malicious code into websites associated with at least 178 organizations, including CNN, Zoom and Equifax.
Read More on News
Users also reported issues with services including Gmail, Apple Pay and Marriott, while the National Security Agency (NSA) confirmed that some of its internal systems were disrupted during scheduled routine maintenance on Tuesday.

Questions remained about the series of disruptions by Tuesday evening.
Although technical problems have been identified as the causes of several incidents, the full circumstances of others remain unclear. No evidence has emerged linking the outages to ARTEX.
FBI Issued Warning About ARTEX AI System
The FBI described ARTEX 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, from reconnaissance through data exfiltration.”
Hackers have been able to use the system to target organizations across numerous sectors including “financial services, government, technology, defense, semiconductor, energy, healthcare, manufacturing, telecommunications, and legal industries,” according to the advisory.
ARTEX reduces the time required to chain attack techniques, according to the advisory. Some intrusions have progressed “from initial targeting to privileged access” in fewer than three hours.
“The framework enables threat actors with minimal technical expertise to conduct complex, multitarget operations simultaneously, fundamentally accelerating attack velocity and expanding operational scale,” the advisory reads.
It has been used to target financial institutions in South Korea, resulting in the theft of personal information relating to approximately 67,000 people, and to compromise a supply chain affecting ConnectWise Cloud Backup customers, enabling access to more than 2,250 downstream tenants.
It was first released onto GitHub in July 2026 and gained “rapid adoption.” The developer said the tool was meant for “authorized security testing within defined scopes” but removed it on October 8 after it became known it was being used maliciously. However, the FBI said it continues to “proliferate” online.
Researchers had identified more than 1,600 publicly exposed ARTEX console instances by early November, the FBI said.
How ARTEX Attacks Work
Artex is one of the first concrete examples of AI-enhanced hacking attack software, George Berg, professor of cybersecurity at the University of Albany, told Newsweek.
The system is built on top of common LLMs. It allows “an attacker to write a high-level human language (e.g., English) description of what organization it wants ARTEX to attack and how.”
“ARTEX uses the LLM to translate this into the technical details of how to conduct the attack. It uses agentic AI very effectively. Agentic AI is where LLMs can create a multitude of separate software entities—agents—to do its bidding,” Berg said.
These agents can access websites and servers to “forge ahead and speedily implement the attack the LLM has planned out based on the human’s prompt,” he added.
Artex is able to “quickly explore and exploit weaknesses in its target organization,” and these attacks often use “a lot of agents,” Berg said. He compared it to a “small army simultaneously exploring the cracks and weaknesses of the organization.”
“This allows an ARTEX attack to very quickly find and exploit even relatively obscure weaknesses in an organization or its online partners (vendors, cloud service providers, etc.),” Berg said.
He added that these attacks are “noisy”; agents are “in effect ‘trying all of the locks’” to gain access to an organization’s systems. While these attacks are “not subtle or stealthy at all,” the speed of the agents allow them to overcome that.
“The time window of effective response is very small. By the time an organization’s security team sees the alarms go off, the attack may very well be successful, and multiplying into successive systems in the network,” Berg said.
The speed of ARTEX attacks makes them “particularly concerning,” Anton Dahbura, executive director of the Johns Hopkins University Information Security Institute, told Newsweek.
“Activities that once required skilled hackers working for hours or days can potentially be carried out automatically in a fraction of that time,” Dahbura said.
“An attack might involve rapidly scanning hundreds of systems for weaknesses, exploiting vulnerable applications or stolen credentials, and moving through connected networks before defenders have time to react.”
AI is “changing the economics of cybersecurity by making sophisticated attacks faster, cheaper, and accessible to people with considerably less technical expertise,” Dahbura said.
FBI Addresses ARTEX Advisory
On Tuesday, the FBI clarified in a post on X that the advisory was published last week as it spread across social media following a post from its Dallas office sharing it.
“We released this advisory last week because as we head to the end of the year, it’s important for Americans to protect themselves against criminals and state actors using SI to run familiar techniques at machine speed,” the post reads. “We’ve seen cyber actors use ARTEX to conduct campaigns targeting organizations globally across multiple sectors.”
The FBI uses the term “super intelligence,” or SI, in its advisory.
The post from FBI Dallas had been viewed more than 5 million times and reshared 12,000 times on X.
An FBI spokesperson pointed to the post when reached by Newsweek for comment.
Four ARTEX Risks Raised by FBI
In the latest advisory, the FBI pointed to four potential uses of the system.
One risk is that AI agents can provide automation for time-consuming phases of enumeration and exploitation, allowing smaller groups to run concurrent campaigns “at a scale previously unavailable to them.”
It can identify and exploit “peripheral systems receiving less security attention than flagship applications.” This can include partner and supplier portals, employee applications, mobile backends, sales-support systems and legacy services.
ARTEX agents have also demonstrated the capability to “identify and exploit supply chain attack vectors, including modifying iOS application distribution infrastructure via exposed cloud credentials.”
The advisory reads: “Autonomous agents can systematically map trust relationships between organizations. By identifying and compromising weaker third-party vendors, CI/CD pipelines, or integrated software providers, attackers can leverage ARTEX to pivot from a peripheral supplier directly into high-value upstream enterprise networks.”
Lastly, it can also lower technical barriers to conduct complex intrusions, allowing actors with less expertise to “execute sophisticated campaigns.”
How to Prevent ARTEX Attacks
Dahbura said that individuals should stick to fundamental cybersecurity advice—keeping software updated, using strong and unique passwords, using multifactor authentication and remaining vigilant about suspicious messages.
“Organizations face a more substantial challenge because they must continuously identify vulnerabilities, secure their internet-facing systems, monitor unusual activity, and ensure that automated defenses can respond quickly enough to increasingly automated attacks,” Dahbura said.
The FBI also issued several recommendations to improve security to mitigate the likelihood of ARTEX attacks.
Organizations can reduce their “external attack surface” by taking inventory of internet-facing applications, restricting access to cloud administration and device management, and disabling basic publishing authentication where it is not required.
The FBI also recommends considering “deployment models such as run-from-package that reduce direct modification of deployed application files” and establishing governance around public IP allocations.
It also suggests taking steps to harden identity and access to systems and strengthen application protections. The FBI urged organizations to preserve evidence, including Microsoft Entra sign-in and audit records, Microsoft Purview Audit records, SIEM data and previously configured Microsoft Graph telemetry.
The FBI also recommends that authorized responders remove confirmed unauthorized credentials, permissions and identities after preserving evidence.
Berg said it is unlikely any of the recent outages are connected to ARTEX.
“More likely, we’re just seeing several high-profile examples of just how hard it is to get complex software and hardware systems right. Even before the advent of AI attacks, computer systems routinely failed. The truth is that really complex systems like the ones we’ve seen at the FAA, the airlines, Apple, the NSA, etc. are fragile,” Berg said.
Dahbura said it is important not to confuse a coincidence with causation
“While ARTEX certainly demonstrates how AI can amplify cyber threats, determining whether it played any role in a particular outage requires forensic evidence, not speculation,” he said.
Director of National Intelligence Jay Clayton told CNBC the outages are not a cyberattack or a “coordinated effort by a nation-state.”
“Many people are patching their systems. They know that their legacy systems have vulnerabilities. They are outdated. They don’t work well with new systems. It is time for an upgrade across our digital infrastructure,” Clayton said.
“As you patch and upgrade your systems, taking them down and bringing them back up, there are things like this that happen.”
Web Outages Fuel Concern
No public evidence has drawn any link between these outages. There has also been no indication from officials that the outages are cyberattacks or related to AI.
The incidents for which causes have been identified appear to involve separate technical failures.
The FAA disruption was linked to faulty networking equipment, while Microsoft attributed its outage to a technical issue. The NSA said its internal systems experienced disruption during scheduled routine maintenance, although it did not explain precisely what went wrong.
There are growing calls for answers pertaining to the outages. Conservative activist Laura Loomer, a close ally to Trump, wrote in a post on X that “obviously something is very wrong.”
House Minority Leader Hakeem Jeffries, a New York Democrat, said during a press conference on Tuesday the outages are “troubling.”
“It’s unclear if it’s all infrastructure related or whether there’s some malignant activity. It’s too premature to know in the absence of a high level briefing for the gang of eight,” he said.
He said there is need for a “bicameral briefing” for the House and Senate about “all of these developments that have taken place in the last week or so.”
“I think it’s time for the administration to come up to the Hill and explain themselves so we can be informed about the steps that need to be taken and we can inform people of the United States of America,” he said.
President Donald Trump downplayed the outages, telling reporters, “It’s life” and that these events are to be expected “when you’re big and powerful and you have all the money.”
Homeland Security Secretary Markwayne Mullin suggested during an interview on Fox News Monday evening that AI may have played a role in some of the disruptions.
“These programs aren’t designed to work as fast as [artificial intelligence] is. And so when they put in a program that’s able to write codes, and it goes in and it self-fixes itself, as it starts self-fixing itself, it’s like doing a change order on a house. When you change one thing in a house, everything else starts changing. That’s a ripple effect,” he told host Sean Hannity.
Newsweek’s reporters and editors used Martyn, our AI assistant, to produce this story. Learn more about Martyn here.
Contact Newsweek editors on this story: Samantha Beech, James Debens and Rhian Morgan.
Update, 11/18/26 10:39 a.m. ET: this article was updated with expert comment and quotes from Director of National Intelligence Jay Clayton.