China Uses Artificial Intelligence to Detect Cryptocurrency Money Laundering
According to Sedaye Sama News Agency, researchers at the National Police University of China, operating under the Ministry of Public Security, have developed an artificial intelligence framework based on Large Language Models (LLMs) capable of detecting money laundering and financial crimes involving virtual assets such as Bitcoin.
The system combines AI-powered pattern recognition with the analytical capabilities of LLMs to examine transaction flows across blockchain networks. According to the South China Morning Post, the platform has achieved an accuracy rate of around 90% in identifying suspicious transactions.
The system is designed to detect abnormal behavior and patterns associated with common money laundering techniques, including transaction layering, structuring payments into smaller amounts to evade financial monitoring, and other methods used to conceal the origin of illicit funds.
Unlike traditional rule-based monitoring systems, the new algorithm can process vast amounts of on-chain data and identify complex behavioral patterns that conventional methods may overlook.
According to China’s Supreme People’s Procuratorate, 3,259 individuals were prosecuted in 2025 on suspicion of involvement in money laundering linked to virtual assets and underground financial networks. The figure highlights the growing scale of cryptocurrency-related financial crimes in China, despite the country’s strict ban on cryptocurrency trading.
The development reflects China’s dual strategy of restricting cryptocurrency activities while making significant investments in blockchain analytics and digital asset tracking technologies. This approach aligns with a broader global trend in which regulators and law enforcement agencies increasingly rely on advanced AI tools to combat financial crime.
Researchers noted that integrating LLMs provides a significant advantage because the models can analyze not only numerical data but also unstructured information, such as transaction descriptions and communication patterns, offering investigators a more comprehensive picture of suspicious activities. This capability can facilitate investigations and judicial proceedings.
However, experts emphasized that 90% accuracy does not eliminate the possibility of false positives. Therefore, all AI-generated alerts must be verified through human review and legal procedures. The technology is intended to support, rather than replace, judicial oversight and law enforcement decision-making.
According to Crypto News, the technology could serve as a model for other countries developing intelligent systems to combat cryptocurrency-related money laundering as digital assets continue to gain wider adoption.
