Japan to Launch AI-Powered Penetration Testing Team at IPA in November
Ministry of Economy, Trade and Industry will establish a 'white hacker' team to proactively discover system vulnerabilities using AI-driven simulated attacks.
TOKYO — Japan’s Ministry of Economy, Trade and Industry will establish an AI-powered penetration testing team at the Information-technology Promotion Agency in November to discover system vulnerabilities before attackers exploit them, Minister Akazawa announced October 23.
The team will use high-performance AI to conduct simulated attacks against government and industry systems, identifying weaknesses before hostile actors can. Akazawa characterized the unit as “off-white hackers” — specialists who think like attackers but work for defense. The ministry framed the initiative as a structural response to a core cybersecurity problem: defense is reactive, must cover all surfaces, and faces adversaries who probe for single points of failure.
The announcement followed disclosure of a critical vulnerability in the PHP programming language on October 21 by OpenAI and Google subsidiaries. METI coordinated with the National Center of Incident Readiness and Strategy for Cybersecurity to issue alerts through roughly 1,000 industry associations. Several government-affiliated organizations, including Nexi (formerly Japan Bank for International Cooperation) and the High Pressure Gas Safety Institute, temporarily suspended websites to apply security patches.
The move signals a shift from purely defensive postures to offensive reconnaissance in Japan’s national cybersecurity strategy, using the same AI tools attackers deploy. The PHP incident demonstrated how AI can uncover longstanding flaws, and accelerated the ministry’s timeline for deploying its own offensive capabilities.