Use of the AI penetration-testing tool ARTEX has been confirmed in unauthorized access attacks targeting approximately 20 Japanese companies. Eight companies have disclosed information breaches, and 6 additional attack infrastructure servers have been identified.
On 19 October 2026, GMOサイバーセキュリティ byイエラエ announced the results of its investigation into the source of attacks targeting Japanese companies, confirming that ARTEX, an AI-powered autonomous penetration-testing tool, had actually been used. The initial attack infrastructure targeted approximately 20 companies, mainly in Japan, 9 of which were in the same industry.
According to the announcement, 8 of the targeted companies disclosed information breaches between September and October, while 6 additional infrastructure servers associated with attacks targeting Japanese companies were identified. GMO believes the attackers restricted their targets by region and industry and prepared a separate host for each attack. CrowdStrike has also reported that ARTEX was used against financial institutions in South Korea, pointing to the possibility that similar methods were used in Japan. The investigation findings have already been shared with law enforcement agencies.
ARTEX was originally developed as a tool for security assessments, using AI to automate the search for vulnerabilities. However, the tool can be used by anyone, and if attackers employ AI, they can target many companies at the same time. It has been suggested that AI can comprehensively cross-reference vulnerability information for older software versions and may also affect versions that are no longer supported. WAFs alone are not sufficient for protection, and monitoring needs to be strengthened on the assumption that software will be kept up to date.