Cybersecurity company Horizon3 warns that the use of artificial intelligence (AI) tools by cybercriminals is dramatically accelerating the speed and scope of cyberattacks, to the point that some organizations have suffered intrusions that have allowed them to gain privileged access to their systems in less than three hours.

The warning comes after the FBI has published a security alert about ARTEX, an open source framework that cybercriminal groups and actors linked to States are using to deploy multiple AI agents capable of autonomously and simultaneously executing attacks against different organizations.

According to Horizon3, these tools are transforming the cybersecurity landscape by allowing operations that previously required days or weeks of work by experienced attackers to now be completed in a matter of hours. Additionally, once the initial target is established, AI agents can identify vulnerabilities, exploit them, and move laterally through compromised systems with minimal human intervention to prevent cybercriminal attacks.

According to Snehal Antani, CEO of Horizon3, “over the last 30 days we have seen a clear increase in both the volume and speed of cyberattacks.” “What previously required days or weeks of work by an experienced attacker,” says Antani, “can now be done in hours, because tools like ARTEX allow a single person to deploy multiple AI agents capable of planning, exploring, exploiting vulnerabilities and moving laterally through systems with virtually no human intervention.”

Ability to automate operations

Horizon3 finds that ARTEX and similar tools facilitate the large-scale execution of known attack techniques, such as the use of stolen credentials, exploitation of default passwords, misconfigurations, or abuse of permissions in cloud environments. Its main difference lies in the ability to automate these operations and execute them against numerous targets at the same time, multiplying their scope and making it difficult for security teams to respond.

Likewise, Horizon3 warns that the use of practically identical tools and procedures by different cybercriminal groups is complicating the identification of those responsible for the attacks, since their patterns of action are increasingly difficult to distinguish.

For Antani, “the use of ARTEX and other similar frameworks is driving a standardization of attacks that makes their attribution difficult. When groups linked to States and purely criminal organizations use practically the same agents, AI models and procedures, their operational footprints begin to resemble each other and many of the signals that traditionally allowed them to be differentiated are lost.”

Risk of interruptions in digital services

Horizon3 considers that the expansion of these tools, together with the vulnerabilities and security deficiencies that many organizations still accumulate, could translate into an increase in interruptions of essential digital services during the coming months, especially in sectors such as banking, commerce or travel.

In this scenario, the company insists that traditional cybersecurity models, based on periodic vulnerability analyzes or annual penetration tests, are not sufficient to deal with threats that evolve and act at machine speed.

In this sense, Antani points out that “we must hope that this situation becomes the norm and is no longer exceptional. Organizations accumulate a significant cybersecurity debt that tools like ARTEX can easily exploit, so it is foreseeable that during the coming months there will be recurring interruptions in digital services while companies try to quickly improve their resilience.”

For this reason, Horizon3 emphasizes the need to adopt a proactive security approach, based on autonomous and continuous testing that allows us to identify truly exploitable attack paths, verify the effectiveness of corrective measures and anticipate cybercriminals because, as Antani concludes, “security teams that continue to treat risk exposure management as a quarterly analysis or an annual penetration test will not be able to keep up with the attackers. Machine-speed attacks are already a reality and defenses must be able to respond to that same speed.”