Back to News
PolicyAI Understanding briefing

South Korea deploys AI crypto-market surveillance as lawmakers seek wider FIU powers

CoinGeek reports that South Korea’s Financial Supervisory Service has deployed an AI system to flag suspicious crypto trading, while lawmakers propose giving the Financial Intelligence Unit direct investigative powers over unregistered virtual-asset businesses.

By 5 min read
AI-generated editorial illustration accompanying South Korea deploys AI crypto-market surveillance as lawmakers seek wider FIU powers
The short version

CoinGeek reports that South Korea’s Financial Supervisory Service has deployed an AI system to flag suspicious crypto trading, while lawmakers propose giving the Financial Intelligence Unit direct investigative powers over unregistered virtual-asset businesses.

What happened

CoinGeek reports that South Korea’s Financial Supervisory Service has deployed an AI-based system to detect suspected crypto-market manipulation and other unfair trading in real time. The report also says lawmakers have introduced a bill that would expand the Financial Intelligence Unit’s authority to investigate unregistered virtual-asset service providers. The reporting is attributed to Business Korea, AJP News Agency and Yonhap News Agency and has not been independently confirmed here.

CoinGeek reports that the Financial Supervisory Service, South Korea’s financial regulator, announced on August 20 that it had developed and deployed an AI-powered surveillance framework for suspicious crypto trading. According to CoinGeek’s account, which cites local outlet Business Korea, the system is intended to monitor digital-asset activity in real time across markets that operate continuously and involve thousands of tokens traded on domestic and international exchanges. The report does not identify the system’s technical architecture, the exchanges covered, or whether the system is already connected to every data source described.

CoinGeek reports that the framework combines real-time transaction analysis with analysis of publicly available online information and automated deeper reviews of assets considered suspicious. The system is said to flag patterns associated with short-term price manipulation, including dramatic price movements at particular times, price increases involving assets with restricted deposits or withdrawals, and artificial trading-volume inflation such as wash trading. For assets showing unusual price or volume activity, the reported system can examine announcements, news reports, posts and videos, including subtitles and audio, to assess possible false information, pre-trading or encouragement of unfair trading.

According to CoinGeek, generative AI is also used to compile reports summarizing the system’s review. If the Financial Supervisory Service determines that further investigation is warranted, the regulator can request additional trading data from exchanges. CoinGeek describes the new framework as an extension of a trading-analysis platform called VISTA, announced in January. The reported change is that alerts, supporting information and preliminary review are linked in one workflow rather than being handled only at a later investigative stage. The report says the regulator plans to add fund-flow and on-chain tracking features, but it does not say those capabilities are already operational.

Read the primary source: coingeek.com

Why it matters

The reported system would place AI closer to the front of financial-market enforcement by combining anomaly detection, public-information analysis and automated review reports. Its practical value will depend on accuracy, human oversight and whether alerts lead to timely investigations rather than simply increasing the volume of suspected violations.

If CoinGeek’s account is accurate, the deployment represents a meaningful change in how a financial regulator can prioritize suspected misconduct. Crypto markets generate activity around the clock and across multiple venues, making manual review difficult. An automated system could help investigators identify patterns that would be costly or slow to find using staff review alone. That does not establish that the system can reliably distinguish manipulation from legitimate volatility, coordinated but lawful trading, market reactions to news, or unusual activity in thinly traded assets.

The reported use of public posts, announcements, subtitles and audio broadens the system beyond transaction monitoring. That may help investigators compare market movements with public claims or promotional activity surrounding a token. It also introduces additional uncertainty: CoinGeek does not report how the regulator verifies translations, sarcasm, recycled information, anonymous claims or misleading context. Nor does the report provide evidence that the system’s media analysis has produced confirmed cases or improved enforcement outcomes.

The proposed legislative change could matter independently of the AI deployment. CoinGeek reports that lawmakers from South Korea’s People Power Party and 10 other legislators filed a bill on August 20 to amend the Act on Reporting and Use of Specific Financial Transaction Information. The proposal would allow people to report suspected violations to the Financial Intelligence Unit and would authorize the agency to investigate and analyze suspected unregistered operators, or refer matters to investigative authorities when violations are recognized. CoinGeek, citing Yonhap, reports that police suspended investigations or preliminary inquiries into 23 of 25 unregistered virtual-asset service providers referred by the FIU between August 2022 and August 2025. That figure and the reasons for those outcomes have not been independently confirmed here.

What to watch next

Key open questions include the system’s false-positive rate, which exchanges and data sources it covers, how audio and video analysis is validated, and whether lawmakers approve the proposed FIU powers. CoinGeek does not report public performance metrics, enforcement outcomes or the bill’s eventual prospects.

The first practical test is whether the AI system produces actionable leads rather than a larger queue of unverified alerts. CoinGeek does not report precision, recall, false-positive rates, response times, confirmed cases, or examples of enforcement resulting from the new framework. Those measures would be needed to judge whether the system improves oversight. They would also help show whether its benefits come from AI-specific analysis or from simply bringing existing data and investigative steps into a single platform.

The regulator’s planned additions deserve close scrutiny. CoinGeek reports that the Financial Supervisory Service intends to enhance the system with fund-flow and on-chain tracking. The source does not specify which blockchains, wallets, exchanges or jurisdictions would be included, how identity information would be linked to transactions, or what legal process would govern access to additional data. Questions about retention, sharing, auditability and the right to challenge an automated suspicion are also unresolved in the report.

The bill’s progress is another important indicator. CoinGeek reports that it was referred to the Political Affairs Committee on August 21 and was awaiting debate before any National Assembly vote. The source does not establish that the proposal will pass, what amendments may be made, or how the expanded authority would interact with police powers and existing anti-money-laundering obligations. For now, both the AI system’s real-world effectiveness and the legislation’s final scope remain unknown.

Related guides & quizzes

What is AI?AI EthicsAI Models ExplainedFuture of AITest what you know — try a free AI quizLook up an AI term in our glossary
Found this useful?