Chinese police university researchers build an AI system to detect crypto money laundering with about 90% accuracy

Researchers at China's National Police University, which operates under the Ministry of Public Security, have designed a framework based on artificial intelligence and large language models that can identify money laundering and economic crimes involving virtual assets such as bitcoin. The report was published on Wednesday 5 August 2026 (14 Mordad 1405).

The South China Morning Post reported that the system has reached about 90% accuracy in flagging suspicious transactions. By analysing transaction flows across blockchain networks, the algorithm detects abnormal behaviour such as layering and splitting sums into smaller transactions to evade financial supervision. Unlike systems built on predefined rules, it can process large volumes of on chain data.

According to figures from China's Supreme People's Procuratorate, 3,259 people were prosecuted in 2025 on suspicion of laundering linked to virtual assets and underground financial networks. Experts stressed that 90% accuracy does not mean the system is error free, and that its output must be confirmed through human review and legal process.