FSS, strengthens virtual asset investigation functions…AI automatically identifies suspicious groups

FSS, strengthens virtual asset investigation functions…AI automatically identifies suspicious groups

The Financial Supervisory Service (FSS) announced on the 3rd that it has developed its own virtual asset investigation infrastructure incorporating artificial intelligence (AI), strengthening its ability to detect unfair trading.

The FSS developed its own virtual asset investigation infrastructure to compensate for the limitations of responding promptly to unfair virtual asset trading with limited personnel, and this is its second update.

This update is characterized by the advancement of its real-time monitoring system.

It collects and analyzes various information other than price in real-time by utilizing the public application programming interfaces (APIs) of domestic and foreign exchanges. An API is a function that automatically places orders with pre-set conditions in the virtual asset market.

There are a total of 8 exchanges from which information is collected: 5 domestic exchanges including Upbit, Bithumb, Gopax, Coinone, and Korbit, and 3 overseas exchanges including Binance, Coinbase, and OKX.

This function instantly determines whether unfair trading is occurring by comprehensively analyzing virtual asset prices, trading volumes, and abnormal transaction indicators in real-time through a market comprehensive dashboard.

In addition, an AI-based function for extracting suspicious linked groups has also been developed.

Previously, FSS investigators manually checked the linkage between funds and order media to determine suspicious groups, but this function automatically analyzes trading patterns and finds interconnected accounts all at once. It can group accounts exhibiting similar behavior into one group and identify the group containing suspicious accounts as a suspicious group.

The FSS stated, "We will gradually expand our AI-based analysis functions," adding, "In the future, we will develop functions to support the creation of investigation documents using text analysis capabilities, and to present wallets and accounts that require tracking through fund flow analysis functions."