Japan's three largest banks are expected to gain access to Anthropic's Claude Mythos artificial intelligence model as early as the end of May, the Nikkei reported on Wednesday. The banks named in the report are MUFG Bank (TYO:8306), Sumitomo Mitsui Banking Corp (TYO:8316) and Mizuho Bank (TYO:8411).
The Nikkei said U.S. Treasury Secretary Scott Bessent likely informed the banks of the decision during meetings held in Japan on Tuesday. If confirmed, this would represent the first time Japanese institutions have been permitted access to Mythos.
Mythos is described in the report as a high-powered AI model that can identify and exploit software security vulnerabilities far more quickly than previous systems. Access to the model has reportedly been limited to roughly 50 organizations worldwide, a group that has included U.S. companies, American banks and British government agencies.
Tokyo has actively sought access to the model, according to the report, as part of efforts to bolster national cybersecurity. The Nikkei said the Japanese government has been pressing Washington for access while stepping up measures to address potential risks from advanced AI systems.
Japanese Prime Minister Sanae Takaichi instructed cabinet ministers to intensify efforts to discover vulnerabilities in the nation’s infrastructure and to reduce AI-related security risks, the report added. The Nikkei said access to Mythos was likely discussed during Takaichi’s meeting with Bessent on Tuesday.
Context and implications
- Granting Japanese banks access to Mythos would mark an expansion of the small set of organizations worldwide currently able to use the model.
- The move reflects Tokyo’s focus on strengthening cybersecurity amid concerns about the capabilities of cutting-edge AI to both detect and potentially exploit software weaknesses.
- Discussions at high-level meetings in Tokyo, including those involving the U.S. Treasury Secretary and Japan’s prime minister, indicate the matter involved diplomatic and security considerations.