Federal investigators say cybercriminals and state-aligned hackers are using the ARTEX artificial intelligence framework to automate attacks against banks, government agencies, defense contractors and critical infrastructure, with documented breaches exposing tens of thousands of records and compromising access to thousands of corporate accounts.

By yourNEWS Media Newsroom

The FBI has issued a nationwide cybersecurity warning about a powerful artificial intelligence-driven hacking framework that allows cybercriminals and state-aligned attackers to penetrate corporate and government networks at extraordinary speed, with some operations progressing from initial reconnaissance to privileged system access in less than three hours.

The agency’s Nov. 13 FLASH advisory, designated FLASH-20261113-001, identifies ARTEX as an increasingly significant cyber threat capable of automating attacks against financial institutions, government agencies, defense contractors, semiconductor manufacturers, health care providers, energy companies, telecommunications networks and other essential industries.

According to the FBI, malicious actors have used the platform in active campaigns since at least August 2026, with documented attacks involving stolen personal information, compromised cloud services and intrusions conducted through trusted third-party technology providers.

The warning comes as American businesses and government agencies face growing concerns about the reliability and security of interconnected digital infrastructure.

ARTEX was originally developed as an open-source penetration-testing framework, a category of software designed to help cybersecurity professionals identify weaknesses in computer systems.

However, investigators say attackers have repurposed the technology to automate unauthorized intrusions with far less human intervention than traditional hacking operations require.

The framework uses artificial intelligence agents to identify vulnerable systems, test weaknesses, exploit security flaws, obtain elevated privileges and extract sensitive information.

Unlike conventional hacking tools that require operators to carry out individual steps, ARTEX can coordinate multiple tasks after receiving broad instructions from a human operator.

The FBI warned that the technology substantially reduces the time required to conduct complex attacks and allows less experienced hackers to target multiple organizations simultaneously.

The framework can integrate with large language models, including Anthropic’s Claude, OpenAI’s GPT models, DeepSeek, GLM, Grok and Kimi.

ARTEX does not contain its own artificial intelligence model. Instead, it connects external models to specialized software components capable of conducting security testing and exploitation activities.

Researchers examining attacker-controlled installations found that Chinese-developed models were particularly common, with DeepSeek variants frequently used because of their cost and capabilities.

The use of a Chinese-developed AI model does not, by itself, establish that an attack was directed by China’s government.

Nevertheless, federal investigators have connected ARTEX-enabled operations to state-aligned cyber actors.

One of the most significant documented incidents involved a supply-chain compromise affecting customers of ConnectWise Cloud Backup, a technology service used by organizations to manage and protect business data.

According to Microsoft’s investigation published Oct. 21, attackers associated with a China-aligned threat group exploited trusted connections between ConnectWise and Microsoft 365 environments.

The campaign enabled unauthorized access attempts involving more than 2,250 downstream customer tenants.

The attackers used ARTEX to accelerate reconnaissance, identify available permissions and exploit weaknesses in cloud authentication and administrative systems.

Because managed service providers often maintain privileged access to numerous customers, compromising one provider can expose organizations that have not been directly attacked.

The FBI identified that type of supply-chain intrusion as a major concern because it allows attackers to move through established business relationships rather than penetrate each company’s defenses independently.

A separate campaign targeted financial institutions in South Korea.

According to the FBI advisory and cybersecurity research published by CrowdStrike, attackers used ARTEX to automate reconnaissance and exploitation against banking and financial systems.

The operation resulted in the theft of personal information belonging to approximately 67,000 individuals.

Researchers found that the attackers used the framework to discover vulnerabilities, identify accessible databases and retrieve sensitive information.

The operation demonstrated that AI-assisted cyberattacks can move quickly from identifying weaknesses to extracting large amounts of personal data.

The FBI also identified campaigns involving cloud platforms, administrative systems and applications that receive less security attention than organizations’ primary websites or networks.

Such systems include employee portals, supplier platforms, customer databases, mobile application services and older software that remains connected to modern infrastructure.

Attackers can exploit these secondary systems to obtain credentials and move into more sensitive parts of an organization.

The bureau warned that this approach makes businesses vulnerable even when their principal websites and security systems are well protected.

According to BleepingComputer’s examination of the warning, researchers identified more than 1,600 publicly exposed ARTEX management consoles by early November.

The number of installations operating privately or behind corporate firewalls is believed to be substantially higher.

ARTEX first appeared publicly in July 2026 through a GitHub repository maintained by a developer using the name Autumn-27.

The software quickly attracted attention from cybersecurity researchers and penetration testers because of its ability to combine artificial intelligence with automated testing capabilities.

After researchers documented its use by malicious actors, the original developer removed the public repository Oct. 8.

However, the FBI said the removal did not prevent continued distribution.

Copies of the software remain accessible through archived repositories, modified versions and a commercial edition known as ARTEX PRO.

The continued availability of the framework means organizations cannot rely on the original repository’s removal to reduce the threat.

The FBI’s advisory describes a broad collection of techniques used in observed ARTEX-enabled intrusions.

Attackers have exploited improperly secured cloud credentials, vulnerable web applications, weak administrative passwords and trusted connections between corporate networks.

Once inside a system, they have created unauthorized accounts, modified permissions, installed remote-access software and attempted to obtain credentials that provide control over additional machines.

Some operations have involved extensive automated examination of internal networks, including large numbers of employee accounts, databases and administrative systems.

The framework can also accelerate data theft by identifying sensitive records and preparing them for transfer to attacker-controlled infrastructure.

The speed of these operations presents an additional challenge for defenders.

Traditional security monitoring systems may detect suspicious activity, but a rapidly automated attack can achieve its objective before security personnel investigate and respond.

Researchers have also observed ARTEX-generated activity producing substantial volumes of network traffic, including repeated authentication attempts and automated requests.

Those patterns may provide warning signs, although the FBI cautioned that individual indicators should be evaluated alongside other evidence before organizations conclude that a compromise has occurred.

The bureau’s warning was published amid a period of heightened attention to major technology disruptions in the United States.

In recent days, outages involving Microsoft services, the Federal Aviation Administration, financial institutions, United Airlines and Apple have raised public concerns about the vulnerability of essential digital systems.

A classified communications network operated by the National Security Agency also experienced a reported interruption Tuesday.

The ARTEX advisory confirms that sophisticated AI-assisted cyberattacks are occurring against major institutions worldwide. However, federal investigators have not established that ARTEX caused those separate American technology outages or that the disruptions were part of a coordinated campaign.

The distinction is important because software defects, equipment failures and network configuration problems can produce widespread service interruptions without malicious activity.

Nevertheless, the FBI’s findings demonstrate that attackers now possess tools capable of rapidly exploiting weaknesses in the same categories of infrastructure on which businesses, financial institutions and government agencies depend.

The White House has also examined the broader implications of artificial intelligence-driven cyberattacks.

According to a Nov. 13 Reuters report, President Donald Trump’s AI task force discussed threats to critical infrastructure, including banking, energy and emergency services.

The meeting included discussions about cybersecurity reporting, information sharing and the development of standards intended to address risks associated with increasingly capable AI systems.

Federal officials have emphasized that artificial intelligence can strengthen defensive cybersecurity operations while simultaneously enabling attackers to conduct more sophisticated intrusions.

The FBI’s ARTEX advisory reflects that dual-use problem.

The platform was developed for legitimate security assessments, but its capabilities can be redirected toward unauthorized attacks.

To reduce exposure, the bureau urged organizations to identify all internet-accessible applications and restrict access to sensitive administrative systems.

It also recommended reviewing third-party integrations, strengthening identity verification, limiting application permissions and monitoring unusual changes to cloud accounts.

Organizations using Microsoft 365 with ConnectWise integrations were specifically advised to examine authentication records and administrative activity for indications of unauthorized access.

The FBI also encouraged companies to monitor abnormal volumes of requests against employee and customer portals, unexpected credential changes and unusual transfers of sensitive information.

Businesses that identify suspected ARTEX-related activity were instructed to preserve logs and other evidence while taking appropriate steps to contain the intrusion.

The advisory provides technical indicators associated with known campaigns, including network addresses, suspicious infrastructure and behavioral patterns that cybersecurity teams can use during investigations.

The bureau cautioned that some network addresses may subsequently be reassigned to legitimate users, making corroboration important before blocking traffic or attributing activity.

Organizations are encouraged to report suspected attacks through the FBI’s Internet Crime Complaint Center or their local FBI field office.

The latest warning reflects an evolving cybersecurity environment in which increasingly autonomous software can perform tasks that previously required substantial technical expertise and time.

A successful attack that once required extensive manual reconnaissance may now progress through multiple stages in hours, increasing the pressure on organizations to detect and contain intrusions quickly.

The documented theft of information belonging to approximately 67,000 people and the attempted exploitation of more than 2,250 Microsoft 365 customer environments illustrate the potential scale of the threat.

As of Nov. 17, the FBI had not announced the total number of American organizations compromised through ARTEX or established that the framework was responsible for the recent nationwide technology disruptions.

The advisory nevertheless confirms that cybercriminals and state-aligned actors are actively using artificial intelligence to accelerate attacks against sensitive industries, creating additional security challenges for businesses and critical infrastructure operators in the United States and abroad.

Original article