Financial System Report Annex "Use and Risk Management of Generative AI by Japanese Financial Institutions"
The Bank of Japan released a report examining generative AI use and risks among 150 financial institutions, highlighting expansion into core operations and the need for robust risk management frameworks.
Generative AI adoption in Japan’s financial sector is accelerating, moving beyond administrative tasks to core operations that handle customer data. The Bank of Japan’s third survey since fiscal 2024, covering 150 institutions, found institutions anticipate further expansion, including in high-stakes areas. However, risks such as information leakage and unpredictable outputs remain critical concerns that require tailored risk management approaches.
The report identifies key challenges, including direct customer-facing GenAI outputs, autonomous AI agents, and frontier AI integration, which demand careful governance. Financial institutions must address these risks to maintain stable operations while leveraging new technologies. Senior management is urged to balance innovation with risk oversight, ensuring adequate resources and frameworks are in place.
The Bank of Japan emphasizes the need for proactive leadership in risk management, particularly as GenAI applications grow more complex. It highlights the importance of ongoing dialogue with institutions to address evolving challenges. The central bank plans to continue monitoring through examinations, seminars, and direct engagement to support safe adoption.
The report also outlines guidelines for reproducing or quoting its contents, requiring prior permission for commercial use and proper attribution for non-commercial purposes. Financial institutions are encouraged to engage with the Bank of Japan’s findings to align their GenAI strategies with regulatory expectations.