Japan Asked U.S. AI Firms for Cybersecurity Access.

Japan’s government asked U.S. AI developers to expand access to models for cybersecurity work, including system-vulnerability inspections. The request follows Tokyo’s push to use AI in cyber defense, but the request’s providers, models, scope and timetable remain unconfirmed.

On Oct. 26, Prime Minister Sanae Takaichi directed ministries to accelerate checks for vulnerabilities in government systems using advanced AI and the expertise of Japan’s AI Safety Institute. She also called for AI-assisted simulated attack tests and monitoring of cyberattacks across supply chains. Takaichi said, “Verifying system vulnerabilities is also essential to preventing damage before it occurs.” The government set a target of compiling an emergency cybersecurity package around the end of the following week.

In May, Finance Minister Satsuki Katayama said, “OpenAI would, in coordination with the U.S. government, provide certain financial institutions in Japan with access to GPT-5.5-Cyber.” In June, Japan’s government and some financial institutions were participating in Anthropic’s Project Glasswing, and Japan had access to Claude Mythos Preview.

An expansion of Project Glasswing would add about 150 organizations across more than 15 countries, with each required to meet its security requirements. GPT-5.5-Cyber is a limited preview for defenders responsible for securing critical infrastructure.