The Financial Supervisory Commission (FSC) on May 20, 2025, released the results of its survey concerning the use of artificial intelligence (AI) by financial institutions and peripheral institutions. Compared to 2024, the adoption rates of both AI and generative AI in the financial industry have increased. The results show that primary objectives for AI adoption are to enhance operational efficiency and productivity, reduce manpower costs, and improve the customer experience.

In April, the FSC conducted a survey covering a total of 383 financial institutions and peripheral institutions. Among them, 126 institutions have already implemented AI, accounting for 33% of the surveyed institutions, up from 29 percent last year. By sector, banks took the lead with an adoption rate of 87%, followed by life insurance companies at 67% and property insurance companies at 45%.

The survey results indicate that the primary purposes for financial institutions adopting AI are to enhance operational efficiency (30%), reduce manpower (18%), and improve customer experience (15%). Overall, the AI applications are mainly concentrated in internal administrative operations, intelligent customer service, and financial crime prevention. The most commonly used AI technologies include Natural Language Processing (NLP) and Large Language Models (LLMs) (31%), Machine Learning (28%), and Robotic Process Automation (RPA) (23%).

According to the survey results, 61 financial institutions have currently adopted generative AI, accounting for 48% of those using AI—an increase of 21 percentage points from the previous year. The most common generative AI application areas are internal administrative operations (39%) and intelligent customer service (15%). Respondents also reported key challenges including the instability or inaccuracy of generative AI outputs, concerns over data security and privacy protection, and compliance-related issues.

Looking ahead, 47% of financial institutions surveyed (179 institutions) intend to adopt or expand the use of AI. The key focus areas remain internal administrative operations, intelligent customer service, and combating financial crime. Furthermore, financial institutions identified several AI areas suitable for joint development through industry collaboration, including anti-fraud AI technologies, localized financial-sector LLMs, and risk management applications.

The FSC stated that among financial institutions that have already adopted AI, the primary application areas are internal administrative operations and intelligent customer service. The level of automated decision-making utilized by AI systems remains low, with 43% of institutions having no automated decision-making and 40% having automated decision-making less than 25%. This indicates that the financial industry is not overly reliant on AI for decision-making. Moreover, 84% of these institutions reported regularly or irregularly monitoring for potential performance drift. Regarding the future planned collaborations among institutions for AI anti-fraud and LLM development, the FSC takes a favorable view and has initiated cooperation through the FinTech Industry Alliance as it continues to encourage financial institutions to engage in joint technological research and pilot projects.

The FSC will continue to monitor AI adoption in the financial industry and urges financial institutions to actively invest in technological innovation to enhance the efficiency, quality, and competitiveness of financial services while safeguarding consumer rights and ensuring financial market order.

Attachments:
1.    Summary of FSC Survey Results on AI Applications by Financial Institutions and Peripheral Institutions
2.    Overview of AI Applications Publicly Disclosed by Financial Institutions