(April 23): Taiwan is launching a project to develop a large language model for its finance sector, seeking to strengthen its local firms and bypass the limitations of global AI platforms that often lack the nuance of domestic regulations and market practices.
The financial large language model project involves 16 Taiwanese financial institutions. The banks have set up a task force under the Taiwan FinTech Alliance to carry out the initiative.
The group aims to build an AI infrastructure “best suited to Taiwan’s financial industry”, according to a statement released late Wednesday.
The project will be led by CTBC Financial Holding Co and overseen by the Financial Supervisory Commission. The envisioned LLM is based on an open-source model but not Chinese platforms.
The model will be trained on specialised datasets including industry-specific data and regulatory frameworks provided by the FSC. It will use the Digital Ministry’s Sovereign AI database to ensure the model reflects local linguistic and cultural contexts.
The project carries an estimated price tag of NT$40 million (US$1.3 million or $1.62 million) to NT$70 million. Training is scheduled to begin in May, with a prototype slated for the third quarter and a final version expected by year-end.
See also: UBS faces US$20 bil hike in capital needs as Swiss split grows
The move comes as global banking giants such as JPMorgan Chase & Co and DBS Group Holdings Ltd reap the rewards of early AI adoption.
CTBC Bank chairman James Chen said that a divide is emerging between “AI-ready” companies and laggards, with the former seeing higher revenue growth and market valuations.
As the wave of generative AI sweeps across the globe, general-purpose models often face challenges in financial applications, including insufficient localisation and difficulty aligning with domestic knowledge, according to the statement from the Taiwan FinTech Alliance.
See also: Middle East conflict clouds Thai banks outlook as profits slip
The initial phase will focus on banking services, with plans to eventually include the insurance and securities sectors. The alliance expects to debut a dedicated FinLLM-based AI agent in the first quarter of 2027 to assist with complex financial tasks and customer service requirements.
Uploaded by Chng Shear Lane
