Crypto: South Korean Watchdog Expands AI Systems To Track Crypto Manipulation

The Financial Supervisory Service said automated models now scan crypto trading activity across timeframes, reducing reliance on manual investigations.

South Korea’s financial watchdog has expanded the use of artificial intelligence to monitor digital asset markets, signaling a shift toward automated, data-driven enforcement as trading activity grows more complex.

On Monday, the country’s Financial Supervisory Service (FSS) said it has upgraded its Virtual Assets Intelligence System for Trading Analysis (VISTA), a model used to investigate unfair crypto trading.

An automated detection algorithm has been added, the FSS said, which can spot potential price-manipulation periods without manual intervention.

The FSS said the new algorithm uses a sliding-window grid search technique to examine every possible sub-period within a trading dataset. This approach is meant to enable exhaustive analysis of potential manipulation windows that investigators previously had to identify manually.

The watchdog said that performance tests on completed investigation cases showed that the system detected all previously identified manipulation periods, while also flagging additional suspicious intervals that were difficult to spot with traditional analysis.

The FSS also said it secured a 170 million won ($116,000) budget for 2026 for further AI performance upgrades, with additional capabilities to be rolled out in different stages through the end of 2026.

Planned upgrades include tools to automatically identify networks of coordinated trading accounts, analyze abnormal trading-related text across thousands of crypto assets and trace the origin of funds used in manipulation.

Related: South Korea’s Coinone weighs stake sale amid Coinbase speculation

The AI surveillance system upgrade aligns with the regulator's aim to enhance enforcement capabilities in the crypto space.

Source: CoinTelegraph