Published: · Last updated: with original publication · Region: Global · Category: cyber

Source-reported · Independent verification status is not recorded for this article; claims remain attributed to their sources.

FBI Warns AI Platform ‘ARTEX’ Is Powering Faster, Smarter Global Cyberattacks

The FBI says advanced artificial intelligence is now helping hackers scan for weaknesses, break into systems, and steal data at high speed using a platform known as ARTEX. The warning signals a shift where corporate networks, public institutions, and critical infrastructure face AI-accelerated attacks rather than traditional human-paced intrusions.

Cyber defenders now have to assume that some of the attacks hitting their networks are being designed and driven by machines. The FBI has issued an alert that advanced artificial intelligence is being used to execute cyberattacks against companies and institutions around the world, citing a platform called ARTEX that can rapidly identify vulnerabilities, penetrate systems and exfiltrate data.

According to the bureau, ARTEX has been in use by hacker groups since at least August 2026. The system is described as an AI tool able to scan networks for weak points, automate infiltration steps that human operators used to script manually, and pull out valuable information at high speed once inside. Some of the attacks tied to ARTEX have already succeeded in obtaining data, though U.S. officials have not yet detailed which sectors or countries have been hit.

The warning matters because it shifts the balance of effort. Companies and public agencies have long faced human adversaries using off-the-shelf tools and known malware strains. Now the FBI is telling them they are up against an adaptive platform that can do reconnaissance, testing and exploitation at a pace no human red team can match. Where a manual intrusion might have taken weeks of probing, an AI-assisted campaign can cycle through thousands of attack paths in hours.

For IT teams, that means more than just adding another line to a risk register. Security operations centers must expect spikes of highly customized phishing, rapidly mutating malware, and attacks that pivot in real time when defenders close one path. Small and mid-sized firms, which often lack dedicated threat-hunting teams, are especially exposed: the same AI that helps big tech companies monitor their clouds can now be rented or stolen to help criminals go after regional banks, hospitals or manufacturers.

The FBI’s alert lands in an environment already shaped by another warning from the private sector. Google has reported that Russian-speaking actors pulled off what it calls the first-ever fully automated AI cyberattack, suggesting that offensive automation is not a future scenario but a current capability in the hands of at least some threat groups. Together, the two disclosures point to a race in which sophisticated adversaries are learning to string together AI modules—like ARTEX—for reconnaissance, exploitation and data theft with minimal human supervision.

For critical infrastructure operators running power grids, water plants or transport systems, the implications are uncomfortable. They now have to treat AI not only as a tool to spot anomalies in their own networks but as an engine that can generate new kinds of intrusion patterns they’ve never seen before. Even if core industrial systems are air-gapped or protected by legacy protocols, the business networks around them—procurement, billing, remote maintenance—can be mapped and attacked by AI tools that systematically search supply chains for the weakest link.

The shareable lesson from this week’s warnings is stark: AI doesn’t just amplify defenders; it also supercharges the learning curve of attackers who used to be limited by their own time and skill. Once an offensive model is trained on enough exploits and network layouts, it can explore complex targets in ways that feel more like an automated trading bot than a traditional hacker.

What happens next will depend on several choices. Law enforcement agencies will decide how much technical detail to release about ARTEX and similar systems, which can help defenders but also risks advertising capabilities. Regulators and boards will have to decide whether to treat AI-accelerated attacks as a new category of operational risk requiring mandatory testing and disclosure. And security vendors will be judged on whether they can build and deploy defensive AI that keeps pace with the offensive tools that, as of this summer, are no longer theoretical.

Sources