Chinese Police AI Tracks Bitcoin Laundering at 90%

Researchers from China’s national police academy have developed an AI framework capable of detecting illicit cryptocurrency transactions with nearly 90 percent overall accuracy, according to a study published in the peer-reviewed Chinese journal Journal of Intelligence. The breakthrough, reported by SCMP, represents a significant advancement in China’s digital crime-fighting capabilities.

The Research

The study, conducted by researchers from the People’s Public Security University of China — which is directly affiliated with the Ministry of Public Security — was published in May 2026. Corresponding author Dr Sun Jingchao, a researcher specialising in criminal investigation and cybersecurity, wrote that the study “provides a precise, generalisable and interpretable solution for detecting illicit cryptocurrency transactions.”

The AI framework uses a memory module and a large language model to detect pseudonymous cross-border illegal trading practices. It is designed to flag suspicious on-chain activity despite the pseudonymous and cross-border nature of cryptocurrency networks, and is extensible beyond Bitcoin, positioning it as a broad-spectrum blockchain forensics tool for regulatory authorities.

According to NationPress, Dr Sun also noted that the framework “provides an innovative technological pathway for regulatory authorities to combat illicit cryptocurrency transactions and economic crimes.”

Enforcement Context

The development comes as Chinese authorities intensify their focus on cryptocurrency-linked financial crime. In March 2026, the Supreme People’s Procuratorate reported that prosecutors indicted 3,259 people in 2025 for money laundering involving virtual currencies and underground banks — a 14.1 percent increase in cases and 7.5 percent increase in people indicted compared to the previous year, as documented by AML China.

These figures reveal the scale of the challenge: underground banking networks have migrated to crypto rails at a pace that has outstripped manual investigation capabilities. The new AI framework is designed to substantially reduce the investigative burden on law enforcement as transaction volumes continue to rise.

Regulatory Trajectory

China has maintained a hardline stance against cryptocurrency since September 2021, when the People’s Bank of China and nine other government agencies declared all cryptocurrency-related activities illegal. In February 2026, new regulations extended the crackdown to stablecoins and the tokenisation of real-world assets, as CoinCentral reported. The rules forbid any entity, domestic or foreign, from issuing a yuan-linked stablecoin without government approval.

Despite the ban, criminals continue to exploit Bitcoin, Ethereum, and other virtual assets for scams, gambling, and money laundering. The pseudonymous nature of crypto transfers and the absence of central-authority approval requirements make these assets attractive for illicit activities.

Forensic Capabilities

The new AI research builds on related forensic capabilities that Chinese police have recently disclosed. In June 2026, Chinese police published a rare technical paper in Forensic Science and Technology revealing their forensic toolkit for cryptocurrency investigation, as SCMP reported. The paper describes a three-phase process: evidence collection (device seizure and key extraction), transaction tracing (blockchain analysis), and asset seizure (freezing wallets on exchanges).

According to ChinaForum.net, the paper’s authors — Sun Shengbin of the Wenzhou Public Security Bureau and Lou Yandi of the Zhejiang Provincial Public Security Department’s Criminal Investigation Corps — outline methods that include tools from Meiya Pico and platforms Pinghang PF and Pinghang X, with major exchanges like Binance, OKX, and HTX referenced in transaction-tracing methodology.

Analysis: The Interpretability Advantage

What distinguishes the Chinese framework from commercial alternatives is its emphasis on interpretability. Unlike black-box commercial products, a model that can explain its outputs can survive cross-examination in court, making it operationally superior for legal proceedings. This gives the Chinese framework a practical advantage in prosecution contexts.

The research also reflects a broader pattern in which China’s security apparatus is integrating large language model and graph-analysis techniques into financial crime investigation. As China continues to tighten oversight of virtual asset flows, the technology could be adopted by prosecutors and financial regulators as a frontline screening tool.

What’s Next

The study’s authors indicate the framework is designed to be generalisable, suggesting potential deployment across multiple cryptocurrency networks beyond Bitcoin. The trajectory of indictments — 3,259 in a single year — points to sustained enforcement pressure that tools like this are built to support.

As the global race between state actors and private blockchain-analytics firms intensifies, China’s institutional approach — building forensic AI explicitly for state law enforcement — raises important questions about evidentiary standards for crypto crime prosecution and the future of digital financial surveillance.