AJP Asia Intelligence Agency

FIU Allocates 8.1 Billion Won for AI Analysis, Faces Tracking Challenges


The Financial Intelligence Unit (FIU) has encountered limitations in tracking the status of financial transaction information provided to investigative agencies on an annual basis. More than half of the responses from nine law enforcement agencies, excluding the National Tax Service (NTS), remain marked as 'in process.' With 8.1 billion won allocated for the establishment of an artificial intelligence (AI) analysis system next year, experts emphasize the need to enhance post-management systems to reflect the accuracy of information utilization results.

According to a report submitted by Kwak Hyun-jun, the chief expert of the National Assembly's Political Affairs Committee, the number of response targets from the nine law enforcement agencies, excluding the NTS, totaled 7,751 cases from January 2018 to August this year. Of these, only 3,766 cases, or 48.6%, were completed. The remaining 3,985 cases (51.4%) are still pending.

The FIU analyzes suspicious transactions reported by financial institutions to assist law enforcement and tax authorities. It subsequently receives feedback on how the information was utilized to improve its analytical work.

The overall completion rate across all agencies was relatively high. During the same period, out of a total of 218,783 response targets, 192,417 cases were completed, resulting in an 87.9% completion rate.

However, there was significant variation among agencies. The NTS accounted for 96.2% of all information provided, achieving a completion rate of 89.4%. In contrast, the completion rate for the remaining agencies fell below half, highlighting a disparity in performance.

A key issue is the difficulty in specifically identifying the status of information marked as 'in process.' It is necessary to distinguish between cases currently under investigation and those that have concluded but have not yet provided feedback. However, the FIU has been unable to present annual processing status due to limitations in statistical extraction.

Relying solely on cumulative statistics makes it challenging to determine how long information remains in processing. It is difficult to differentiate between recently provided information currently under investigation and older information that has not yet yielded results, creating limitations in assessing response delays.

These post-management issues are intertwined with the FIU's initiative to enhance AI analysis. The FIU announced in February its plans to incorporate AI into its review and analysis system, with a budget of 8.1 billion won allocated for the development of related systems in next year's budget. The goal is to improve the accuracy of AI analysis by incorporating the results of information utilization by law enforcement agencies.

Even with the establishment of an AI system, without a systematic approach to securing how analytical information is utilized in actual investigations and tax assessments, there may be constraints in verifying and improving analytical results. This underscores the need for a system to track the status of information processing and utilization results after information is provided.

The total budget for the FIU's IT network construction and operation project, including the AI system development costs, is set at 14.662 billion won, an increase of 9.861 billion won (205.4%) from this year's budget of 4.81 billion won.

Kwak emphasized the necessity to reform the post-management system to enable time-series tracking and to establish "institutional and procedural incentives or inter-agency agreements" to encourage prompt responses from key investigative agencies.




* This article has been translated by AI.