By Hamza Faisal
Around 1,000 servers suspected of serving as bases for advanced artificial intelligence-driven cyberattacks have been online since October, a Japanese-led international investigation has found.
The probe by the Japan Computer Emergency Response Team Coordination Center (JPCERT/CC), conducted with overseas cybersecurity firms, found the servers may have used AI tools to search for vulnerabilities and target websites around the world. JPCERT/CC began its investigation in early October and is still tracking 600 to 700 of the servers.
An AI toolkit for probing websites
Most of the servers hosted AI-powered platforms, including Artex, a penetration-testing tool designed to automate security checks. Investigators said the servers appeared to be scanning for weaknesses in websites and trying to steal configuration files, the settings that can expose how a system is built and where its defences are weakest.
One server also contained traces of a programming-support tool created by the US AI company Anthropic. Investigators did not conclude that the tool powered the suspected activity, and the finding does not establish how many organisations, if any, were actually breached.
600 servers still being tracked
The scale of the discovery is about suspected infrastructure, not confirmed damage. The roughly 1,000 servers represent machines investigators believe may have been used to launch attacks, not 1,000 successful break-ins, and it has not been established that all of them were physically located in Japan or were originally legitimate systems later compromised.
Still, the fact that JPCERT/CC is actively monitoring 600 to 700 of them weeks after the probe began underlines how persistent the operation appears to be. The servers had been online since October, giving their operators weeks to probe targets before investigators mapped the network.
A warning that worse is coming
JPCERT/CC analyst Ken Sajo warned that AI is still in the early stages of being used in cyberattacks, and that more sophisticated attacks will increase as criminals become more adept at exploiting AI capabilities.
The findings land as Japan steps up its defences. On November 6, the Japanese government announced an emergency cybersecurity package backed by an initial allocation of about 10 billion yen, roughly $63 million, after a series of attacks affecting businesses and government organisations.
For defenders, the probe is a reminder that AI is cutting both ways: the same tools that help security teams find flaws faster are now being assembled into industrial-scale attack infrastructure. Patching promptly, locking down configuration files and watching for automated probing have never mattered more.



