Approximately 1,000 servers suspected of being cyberattack infrastructure using Chinese-made AI were identified worldwide. An international investigation found cases in which automated attack tools were used to attempt unauthorized access to websites in Japan and overseas.
A joint investigation by Japanese information-security organizations and overseas companies has found that at least approximately 1,000 servers believed to be cyberattack infrastructure using sophisticated AI have existed on the global internet since October. The investigation was conducted by organizations including the JPCERT Coordination Center (JPCERT/CC), and includes servers located overseas.
These servers are believed to have targeted websites and web applications in Japan and overseas, using AI to search for vulnerabilities and attempt unauthorized access. The high degree of automation has lowered the barrier to carrying out attacks, creating conditions in which even inexperienced attackers can execute them.
The tools confirmed to have been misused include Cairn, ARTEX and CyberStrikeAI. Signs were found that AI agents had been incorporated into application analysis and vulnerability-testing procedures. Since September 2026, unauthorized-access damage has also increasingly come to light in Japan.