The Financial Supervisory Service (FSS) is accelerating its digital transformation by introducing artificial intelligence (AI) to handle civil petitions and disputes starting this month.
According to financial authorities today (Sept. 2), the FSS will pilot AI services in departments handling civil petitions and disputes across various financial sectors around mid-September.
The FSS plans to fully launch an AI-based comprehensive financial petition and dispute portal next January.
This is part of the mid-to-long-term digital innovation (DX) project that the FSS has been pursuing since late last year.
The FSS aims to introduce a categorization system by key issues in the petition handling process and improve processing efficiency through AI-based workflows.
AI will be utilized across the entire process, including petition classification, recommendation of similar cases and precedents, and drafting of reply letters.
However, reply letters will be subject to final review and writing by staff members to reduce the possibility of errors.
To achieve this, the FSS has been developing AI services by introducing Naver's AI model HyperCLOVA X and open-source software into its internal network.
With financial petitions on the rise, the introduction of AI is expected to improve processing speeds.
Last year, financial petitions reached 128,419 cases, a 10.4% increase from the previous year.
The number of processed cases also rose by 17.0% to 127,809, but the processing period increased by 5.1 days to 46.6 days.
An FSS official said, "We are pursuing AI-based automation for each task unit," and added, "We will enhance model consistency through ongoing training until the comprehensive portal is fully launched next year."
In addition, the FSS is working on establishing an AI response system to detect unfair trading practices by introducing AI into investigation operations.
The AI for unfair trading investigations is scheduled to be implemented starting in 2028.
(Photo: Yonhap News)
※
Copying, redistribution, and unauthorized use in AI training are strictly prohibited.