Generative AI Use and Risk Management: Fiscal 2026 Survey Results
Overview
The Bank of Japan’s Financial System and Bank Examination Department organized the current state, challenges, and risk management issues surrounding the use of generative AI in the financial industry based on a survey of 150 client financial institutions and discussions with financial institutions, IT vendors, and others. More than 90% of financial institutions are using or testing generative AI, and its use is expanding from general administrative tasks to core operations involving customer information. Meanwhile, there is room for improvement in governance, third-party risk management, safety and security, among other areas. Further expansion of use will require management involvement and the establishment of appropriate organizational frameworks.
Key points
- The use of generative AI is expanding from general administrative tasks to core operations involving customer information.
- Only a small number of financial institutions currently present generative AI outputs directly to customers.
- There is room for improvement in governance, third-party risk management, safety, and security.
- Management involvement and the development of risk management frameworks will be important for expanding the use of generative AI going forward.
Overview
Generative AI is rapidly spreading throughout society, and its use is also expanding among financial institutions in Japan. In addition to general administrative tasks, its applications are extending to core operations that use customer information, and further expansion and sophistication are expected as technological advances continue to be adopted.
The risks specific to the use of generative AI include the risk of information leakage and uncertainty regarding its outputs and behavior. After understanding these risks, financial institutions are expected to establish management frameworks based on a risk-based approach.
Based on the survey results and discussions with financial institutions, IT vendors, and others, this report organizes the current state and challenges of generative AI use, as well as issues related to risk management.
Impact
As the use of generative AI advances, the challenges facing financial institutions and the risks requiring management are expected to become more complex and diverse.
To advance the use of generative AI while maintaining stable business operations, financial institutions are expected to have management understand and lead the assessment of benefits and risks, while developing the necessary resources and organizational frameworks.
Details
The Bank of Japan conducted a survey of 150 client financial institutions. This was the third survey since fiscal 2024.
The survey found that more than 90% of financial institutions were using or testing generative AI, and the share using it increased across all business categories. The increase over the past year was particularly notable among second-tier regional banks.
In evaluations after implementation, many respondents said that generative AI could be used for business operations but that there was room for improvement. The responses also reflected expectations for financial institutions to improve their own capabilities and for generative AI technology to advance further.
Information management and practical rules have been established at a considerable number of financial institutions. Meanwhile, many financial institutions said that there was room for improvement or that they were still considering measures related to governance, third-party risk management, safety, and security.
Regarding responses to challenges, many financial institutions reported progress in improving guidelines and IT infrastructure, while business development, human resources training, IT infrastructure and data development, and cybersecurity measures are recognized as future challenges. For the direct presentation of outputs to customers and the introduction of AI agents and frontier AI, responses in terms of risk management and governance will be important.
The Bank of Japan plans to continue monitoring the use of generative AI through examinations, monitoring, seminars, and other activities, while continuing dialogue with financial institutions on appropriate approaches to risk management and governance.