Taiwan's Dedicated Financial Brain to Launch by Year-End, Industry Optimistic about Efficiency and Talent Development
Taiwan is set to launch its first financial-specific large language model (FinLLM) by the end of the year. With 16 banks participating, the project is expected to significantly boost efficiency in banking operations, such as loan approvals, and foster FinTech talent, leading to a new era for Taiwan's financial services.
📋 Article Processing Timeline
- 📰 Published: April 25, 2026 at 13:23
- 🔍 Collected: April 25, 2026 at 13:31 (8 min after Published)
- 🤖 AI Analyzed: April 25, 2026 at 13:42 (10 min after Collected)
Central News Agency
(Central News Agency reporter Su Ssu-yun, Taipei, 25th) Taiwan is building its first large language model for the financial industry, with the banking version expected to go live by the end of this year. Its capabilities are equivalent to those of an entry-level professional. Industry players anticipate two major benefits from the project: first, the model will be more precise and efficient, for example, accelerating the review of bank loan applications; second, it will drive the cultivation of financial technology talent.
The Financial Technology Industry Alliance, guided by the Financial Supervisory Commission (FSC), recently announced the launch of the Financial Large Language Model (FinLLM) project. This initiative aims to build a financial-specific large language model equipped with local financial knowledge and an understanding of regulatory frameworks. Sixteen banks have invested resources and manpower. Subsequently, the National Chengchi University FinTech Research Center and financial industry peers will establish a model evaluation mechanism, serving as a crucial foundational project for the financial industry's future adoption of AI.
FSC Chairman Peng Chin-Lung pointed out that the banking version of the model will be launched by the end of this year, and the next phase will expand to the insurance and securities sectors.
CTBC Financial Holding serves as the convener for this project. Jia Jing-Guang, CIO of CTBC Financial Holding, noted that the model's data sources include knowledge related to certifications from the Taiwan Academy of Banking and Finance, and all currently effective legal and regulatory content. The model's expected capability is akin to an entry-level bank employee, and its greatest advantage is that it will be the AI model with the deepest understanding of Taiwan's regulations and local financial knowledge.
Regarding the application benefits of concern to the public, Jia Jing-Guang used the analogy of "taking an exam." He explained that banks currently use foreign models, which is like an "open book exam." While AI can be instructed to find answers from specified data to avoid errors, efficiency drops when the data volume is too large.
Jia Jing-Guang further explained, "With the financial large language model in the future, it will be like having a financial brain with built-in knowledge." This will evolve from "passively flipping through books to find answers" to "actively applying built-in knowledge," with a vast difference in efficiency between the two.
CTBC Bank added that in the future, various banks can develop applications based on this model. Its core has a built-in deep understanding of Taiwan's financial regulations and operations, which will fundamentally improve accuracy and credibility. For example, for loan applications, the financial large language model as a foundation will accelerate review, enhance response accuracy, and thereby improve customer experience.
In addition, CTBC Bank observed that another major benefit lies in the cultivation of financial industry technology talent. This project brings together 16 banks, and all participating project personnel will gain deep exposure to core processes such as model data cleaning and training, which is highly valuable for the AI capability development of each bank.
KGI Securities, also a participant in the project, stated that AI is not only about technology adoption but also a comprehensive transformation of financial services and organizational operating models. The group integrates AI development with its ONE KGI one-stop financial services platform, collaborating with its subsidiary KGI Bank, to promote AI implementation through three strategies: intelligent compliance governance, agile innovation incubation, and talent development through collaborative learning.
KGI Securities pointed out that strengthening regulatory analysis and risk monitoring through the capabilities of the financial large language model will shift compliance management from passive review to proactive defense. Utilizing AI to integrate cross-departmental knowledge will accelerate the development process of innovative services and create a generative learning environment, establishing a personalized financial professional development system to cultivate financial talents with AI collaboration capabilities. (Edited by Zhai Sijia) 1150425
Choose to stand with facts, every sponsorship you make is a force for protecting press freedom.
Download the Central News Agency "First-hand News" APP to stay updated with the latest news.
Text, images, and audio/video content on this website may not be reproduced, publicly broadcast, or publicly transmitted and used without authorization.
(Central News Agency reporter Su Ssu-yun, Taipei, 25th) Taiwan is building its first large language model for the financial industry, with the banking version expected to go live by the end of this year. Its capabilities are equivalent to those of an entry-level professional. Industry players anticipate two major benefits from the project: first, the model will be more precise and efficient, for example, accelerating the review of bank loan applications; second, it will drive the cultivation of financial technology talent.
The Financial Technology Industry Alliance, guided by the Financial Supervisory Commission (FSC), recently announced the launch of the Financial Large Language Model (FinLLM) project. This initiative aims to build a financial-specific large language model equipped with local financial knowledge and an understanding of regulatory frameworks. Sixteen banks have invested resources and manpower. Subsequently, the National Chengchi University FinTech Research Center and financial industry peers will establish a model evaluation mechanism, serving as a crucial foundational project for the financial industry's future adoption of AI.
FSC Chairman Peng Chin-Lung pointed out that the banking version of the model will be launched by the end of this year, and the next phase will expand to the insurance and securities sectors.
CTBC Financial Holding serves as the convener for this project. Jia Jing-Guang, CIO of CTBC Financial Holding, noted that the model's data sources include knowledge related to certifications from the Taiwan Academy of Banking and Finance, and all currently effective legal and regulatory content. The model's expected capability is akin to an entry-level bank employee, and its greatest advantage is that it will be the AI model with the deepest understanding of Taiwan's regulations and local financial knowledge.
Regarding the application benefits of concern to the public, Jia Jing-Guang used the analogy of "taking an exam." He explained that banks currently use foreign models, which is like an "open book exam." While AI can be instructed to find answers from specified data to avoid errors, efficiency drops when the data volume is too large.
Jia Jing-Guang further explained, "With the financial large language model in the future, it will be like having a financial brain with built-in knowledge." This will evolve from "passively flipping through books to find answers" to "actively applying built-in knowledge," with a vast difference in efficiency between the two.
CTBC Bank added that in the future, various banks can develop applications based on this model. Its core has a built-in deep understanding of Taiwan's financial regulations and operations, which will fundamentally improve accuracy and credibility. For example, for loan applications, the financial large language model as a foundation will accelerate review, enhance response accuracy, and thereby improve customer experience.
In addition, CTBC Bank observed that another major benefit lies in the cultivation of financial industry technology talent. This project brings together 16 banks, and all participating project personnel will gain deep exposure to core processes such as model data cleaning and training, which is highly valuable for the AI capability development of each bank.
KGI Securities, also a participant in the project, stated that AI is not only about technology adoption but also a comprehensive transformation of financial services and organizational operating models. The group integrates AI development with its ONE KGI one-stop financial services platform, collaborating with its subsidiary KGI Bank, to promote AI implementation through three strategies: intelligent compliance governance, agile innovation incubation, and talent development through collaborative learning.
KGI Securities pointed out that strengthening regulatory analysis and risk monitoring through the capabilities of the financial large language model will shift compliance management from passive review to proactive defense. Utilizing AI to integrate cross-departmental knowledge will accelerate the development process of innovative services and create a generative learning environment, establishing a personalized financial professional development system to cultivate financial talents with AI collaboration capabilities. (Edited by Zhai Sijia) 1150425
Choose to stand with facts, every sponsorship you make is a force for protecting press freedom.
Download the Central News Agency "First-hand News" APP to stay updated with the latest news.
Text, images, and audio/video content on this website may not be reproduced, publicly broadcast, or publicly transmitted and used without authorization.