TrendAI maps three AI-assisted attack infrastructures targeting Japan, including a 44-agent CyberStrikeAI deployment probing small firms and municipalities

Trend Micro's research arm TrendAI published on 27 October the English edition (Japanese edition 23 October) of an analysis that separates AI-assisted attack infrastructure targeting Japan into three strands. In the strand that ran the open-source agentic penetration-testing tool CyberStrikeAI with 44 agents, automating reconnaissance through intrusion, exfiltration and reporting, at least 19 organisations were probed in a short burst in October, at least 16 of them Japanese small businesses and local public bodies; Japanese-language fake login pages were generated, and the deepest intrusion began at an unauthenticated admin interface on a legacy system. The authors caution that the operators are unknown, that the Simplified Chinese console is a weak indicator, and that the findings are no proof of a new state-backed campaign.