South Korea’s FSS to Monitor Crypto Trading in Real Time, Use AI to Detect Borrowed-Name and Collusion Rings

South Korea’s Financial Supervisory Service is upgrading its in-house surveillance system by combining artificial intelligence with real-time monitoring to combat increasingly sophisticated and complex unfair trading in virtual assets.

The centerpiece of the second upgrade to its virtual-asset investigation infrastructure is a real-time monitoring system and an AI-based function that automatically flags suspicious groups, the FSS said on May 3. The regulator developed the system in-house to offset limited investigative manpower and improve surveillance efficiency.

The FSS has built a real-time monitoring system linked to the public application programming interfaces, or APIs, of domestic and overseas crypto exchanges. Moving beyond simple price monitoring, the system collects and analyzes a wide range of market data in real time to detect abnormal trading immediately. The regulator said it plans to identify tokens targeted for unfair trading at an early stage, launch investigations promptly and minimize investor losses.

The agency also expanded its AI-based tracking of suspected groups. After developing an AI algorithm-based function in January that automatically extracts suspicious trading periods, it has now added a function that automatically identifies linked groups suspected of price manipulation.

Used together, the two functions can quickly identify organized suspicious groups, including collusive price-manipulation networks and accounts suspected of being opened under borrowed names. The system can also analyze the trading periods of flagged groups down to the second, improving the speed and accuracy of investigations.

The FSS said it has established a framework to swiftly detect unfair trading in the virtual-asset market and take strict action. It added that it will continue expanding the use of AI to protect users and help establish a sound market order.