The Financial Supervisory Service (FSS) is automating its market monitoring of unfair trading in virtual assets using artificial intelligence (AI).
On August 20, the FSS announced the establishment of an AI-based monitoring system for the virtual asset market that integrates generative AI and machine learning algorithms.
This system includes real-time transaction data analysis, online public information analysis, and automatic in-depth analysis of suspicious assets.
Initially, the AI algorithm will capture abnormal transactions occurring in the virtual asset market in real time. Based on the FSS's accumulated investigative experience, it will automatically identify suspected short-term price manipulation cases, such as the 'racehorse' type, where prices fluctuate dramatically at specific times, and the 'cage' type, where prices surge for assets with restricted deposits and withdrawals.
For assets with significant price and trading volume movements, generative AI will analyze related announcements and news to determine the causes of price fluctuations. If typical price manipulation patterns emerge or prices rise sharply without clear reasons, the FSS will request trading data from exchanges to assess the need for further investigation.
The AI will also detect artificial trading volume inflation, such as wash trading. By applying Benford's Law and machine learning algorithms, it will identify assets and trading periods that deviate from normal trading patterns.
Unfair trading occurring online will also be monitored. Generative AI will analyze posts and videos related to virtual assets, examining subtitles and audio to detect illegal pre-trading through leading rooms, dissemination of false information, and incitement to unfair trading.
After detecting abnormal transactions, AI will be utilized in the in-depth analysis process. It will assess quantitative indicators such as price and trading volume surges, the causes of price increases identified by AI, complaints, tips, and media reports to determine whether to initiate a detailed analysis.
Generative AI will also compile review reports summarizing the analysis results. Based on the indicators generated by AI, the reports will be automatically created in a pre-designed format, allowing investigators to decide on further analysis or planning investigations.
An FSS official stated, "With the establishment of this AI-based market monitoring system, we expect to respond quickly and efficiently to increasingly sophisticated and complex unfair trading in virtual assets with limited personnel. We also plan to further enhance the AI-based market monitoring and investigation system by developing additional features to support fund flow and on-chain tracking."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.