SIAgentNews.com
← All news
DevelopingSecurity

FBI warns criminals are using ARTEX AI to automate cyberattacks

Cybercriminals are repurposing a security testing tool to breach organizations in under three hours, leaving defenders with very little time to react.

· First reported by malwarebytes.com

under three hourstime to reach privileged access

The short version

  • Attackers have used ARTEX to reach privileged access in under three hours in some observed cases.
  • ARTEX is an open-source penetration-testing framework that connects to external large language models.
  • The FBI reports activity against finance, government, healthcare, and technology sectors since at least August.
  • Microsoft identified a China-state-aligned group using the tool to accelerate attacks after stealing credentials.

Why it matters

This shift toward AI-driven automation means that common security gaps can be exploited much faster than human defenders can typically notice. It heightens the risk for individuals whose personal data is held by companies with unpatched vulnerabilities.

The rise of automated intrusion

The FBI recently issued a FLASH advisory to warn that cybercriminals are abusing a tool called ARTEX to automate intrusions against organizations worldwide. This tool was originally developed for authorized security testing, which involves trying to break into systems with the owner's permission to find weaknesses. By repurposing it for malicious use, attackers can execute complex hacks with much less manual effort than before.

In some observed attacks, criminals were able to move from selecting a target to gaining privileged access in under three hours. This rapid pace is a significant concern for security teams because it leaves very little time to investigate or stop the breach. An alert is only truly useful if defenders can act before the damage is already done by the intruder.

As malwarebytes.com reported, the FBI has seen this activity occurring since at least August. The attacks have targeted a wide range of critical sectors, including government, healthcare, finance, and technology. These broad targets suggest that the tool is being used to find and exploit various types of organizational vulnerabilities at scale.

How the ARTEX framework works

ARTEX is an open-source penetration-testing framework that functions as a planning agent for cyberattacks. When an operator gives the system a target and a specific objective, the planning agent divides the work into smaller, manageable tasks. It then coordinates other AI agents to use security tools to carry out those tasks and share their findings with the main system.

One unique technical aspect of ARTEX is that it does not have its own built-in AI model. Instead, it connects to external large language models (LLMs) to power its decision-making and planning capabilities. This allows the framework to leverage the reasoning power of existing high-end AI to navigate complex digital environments and identify paths into a network.

The automation provided by ARTEX helps attackers connect individual weaknesses into a cohesive, larger intrusion. For example, an overlooked employee portal or an exposed cloud credential can be linked together to reach sensitive records. By automating these connections, the AI can find routes to additional accounts or even connected organizations that a human might miss.

Who is at risk and why

While the primary targets of these attacks are large organizations, the consequences can reach individual home users. If a company's records are breached, personal information can be stolen and later used for fraud or targeted scams. This means that the speed of the AI-driven attack directly impacts the safety of everyday people who trust these companies with their data.

Microsoft reported a specific instance where a China-state-aligned group used ARTEX to accelerate their attacks. This group managed to use the tool after they had already compromised credentials for a ConnectWise application. This highlights how attackers can combine stolen credentials with AI automation to move through a network much faster than traditional methods allow.

The FBI notes that while these AI-driven attacks generate plenty of suspicious activity, the sheer speed of the process is the main danger. The automation allows criminals to do more of the hacking for them, reducing the need for highly skilled human operators to stay involved in every step. This lowers the barrier to entry for criminals while increasing the velocity of the threat.

Defending against AI speed

The FBI clarifies that ARTEX does not make basic security controls obsolete, but it does make leaving familiar weaknesses unfixed much more dangerous. Because the AI can find and exploit gaps so quickly, organizations must be more proactive about their maintenance. The goal for businesses is to close the doors before the AI agent can find them and map out a path through the system.

For organizations, the FBI recommends staying up to date by promptly patching internet-facing systems and inventorying overlooked applications. This includes old employee or supplier portals that might be forgotten but still provide a way into the network. IT teams should also restrict administrative interfaces to approved paths and rotate any compromised secrets or API keys.

For individuals and small-business users, the fundamentals of cybersecurity remain the most effective defense. Users should continue to update their devices and software, use strong unique passwords, and enable multi-factor authentication. Running an up-to-date real-time anti-malware solution is also a key step in protecting personal devices from these types of automated threats.

What to watch

  • How other open-source security tools might be repurposed by attackers.
  • The development of AI-specific defense tools to counter automated penetration testing.

Try it

Check your organization's inventory for any old employee or supplier portals that are no longer in use and ensure they are decommissioned.

Sources and links

First reported by malwarebytes.com. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.

Liked it? Say why

Add a line for other readers.

No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.

More AI news