JPCERT confirms about 1,000 attack servers worldwide suspected of abusing AI, probing website weaknesses around the clock
AIが24時間休まず弱点を探す攻撃用サーバー、世界に約1,000台をJPCERTが確認 – shiritomo
JPCERT Coordination Center has been investigating in cooperation with overseas security companies since early October 2026, and as of November 14 confirmed that at least about 1,000 servers worldwide are suspected of being used as attack bases for abusing AI to probe website weaknesses and attempt intrusions; overseas media reported that 600–700 of them are being continuously tracked. Analysis of 3 captured servers showed that one had the APK analysis tools Blutter and jadx installed, used to find undisclosed APIs; one used Python scripts to automatically test for SSRF vulnerabilities; and another targeted web servers, MySQL and Jenkins, with traces of EDR evasion. Overseas media reported that traces of Anthropic's coding assistant tool were also found on the attack servers, possibly repurposed for writing attack code; JPCERT analyst Ken Sajo said AI use in cyberattacks is still at an early stage and methods may become more sophisticated.