South Korea uses AI to catch crypto market manipulation in under an hour

South Korea's financial watchdog is leveraging artificial intelligence to combat illicit activities in the virtual asset market. The Financial Supervisory Service (FSS) announced the operational launch of its automated market surveillance process, integrating generative AI and machine learning algorithms.

This advanced system aims to detect and respond to unfair trading practices, including market manipulation, across approximately 2,000 virtual assets traded on the country's five major won-based exchanges: Upbit, Bithumb, Coinone, Korbit, and Gopax. Previously, the in-depth analysis of transaction data and the creation of review reports were manual processes that could take up to three days. The new AI-driven system slashes this timeframe to less than an hour.

“

Previously, we could only initiate in-depth analysis for cases with relatively clear suspicions, but now we can screen even somewhat ambiguous ones by putting them into the system.

The FSS's Virtual Asset Investigation Bureau stated that the AI algorithms can now identify trading patterns indicative of market manipulation and wash trading. Generative AI further enhances this capability by analyzing news, online posts, and videos to detect illegal front-running schemes, the spread of false information, and incitement to unfair trading through online communities.

Previously, the FSS could only prioritize cases with strong suspicions due to time constraints. However, the AI system allows for the screening of even ambiguous cases, potentially leading to more actual investigations. This technological leap is expected to significantly bolster the FSS's capacity to maintain a fair and transparent virtual asset market.

“

There have been cases where the detection algorithms led to actual investigations.