As digital currencies weave themselves ever deeper into the global financial fabric, the ancient tension between concealment and detection has found a new arena. Chinese law enforcement, confronting a surge in cryptocurrency-linked money laundering, has turned to artificial intelligence — combining memory modules and large language models — to pierce the pseudonymous veil of blockchain transactions with a claimed 90 percent accuracy. The move reflects not merely a technical ambition but a broader reckoning: that the speed and borderlessness of crypto crime now demands tools as adaptive and rel
Chinese police deploy AI to detect bitcoin laundering with 90% accuracy
An innovative technological pathway for regulatory authorities to combat illicit cryptocurrency transactions
Mark
Why does China need a special AI system for this? Couldn't traditional financial crime investigators just track the transactions?
Mimi
Cryptocurrency moves at machine speed across borders. By the time a human analyst finishes one investigation, the money has already moved through five different wallets in three countries. The AI can process thousands of transactions simultaneously and spot patterns humans would never see.
Mark
So the 90 percent accuracy—what does that actually mean? Is it finding 90 percent of all laundering, or 90 percent of the transactions it flags?
Mimi
The research doesn't specify that distinction, which matters. It likely means that when the system flags something as suspicious, it's correct 90 percent of the time. That's useful for investigators, but it doesn't tell us how much actual laundering is still slipping through undetected.
Mark
The numbers from the procuratorate—3,259 indictments in one year. Is that a lot?
Mimi
For a country of 1.4 billion people, it's a meaningful enforcement effort but not enormous. What's notable is that they're tracking it separately, treating crypto crime as its own category. That suggests they see it as a growing problem, not a marginal one.
Mark
Does this technology actually stop the money laundering, or just catch people after the fact?
Mimi
It catches people after the fact. The AI identifies suspicious patterns, investigators follow up, and then prosecution happens. By then the money may already be spent or hidden. But the threat of getting caught changes behavior—if criminals know their transactions are being analyzed, some will stop trying.
Mark
What happens to the people who get caught?
Mimi
The source doesn't say. But given that these are indictments for money laundering, not just possession, we're talking about serious criminal charges. China's penalties for financial crimes can be substantial.
Il Polso
- Cryptocurrency's pseudonymous architecture has given money launderers a structural advantage over traditional banking surveillance, allowing illicit funds to cross borders faster than regulators can respond.
- China's own enforcement data reveals the scale of the crisis — 3,259 people indicted in a single year for virtual currency money laundering, a figure that signals a problem too large for conventional investigative methods alone.
- Researchers affiliated with China's Ministry of Public Security have published an AI framework that maps wallet relationships over time and cross-references contextual data, training itself on historical laundering patterns to flag new ones.
- A 90 percent detection accuracy rate, if it holds in operational conditions, could transform overwhelmed enforcement agencies by letting human analysts concentrate on the highest-value cases rather than sifting raw transaction noise.
- The deployment positions China at the leading edge of a likely global shift, where AI-driven financial crime detection becomes standard practice and the arms race between regulators and crypto criminals enters a new, algorithmic phase.
As digital currencies weave themselves ever deeper into the global financial fabric, the ancient tension between concealment and detection has found a new arena. Chinese law enforcement, confronting a surge in cryptocurrency-linked money laundering, has turned to artificial intelligence — combining memory modules and large language models — to pierce the pseudonymous veil of blockchain transactions with a claimed 90 percent accuracy. The move reflects not merely a technical ambition but a broader reckoning: that the speed and borderlessness of crypto crime now demands tools as adaptive and relentless as the crimes themselves.
China's law enforcement agencies have begun deploying an AI system built to identify bitcoin money laundering, with researchers claiming it correctly flags illicit transactions nine times out of ten. The framework pairs a memory module with a large language model, allowing it to trace cryptocurrency flows across borders and cut through the anonymity that makes blockchain an attractive vehicle for moving dirty money.
The problem the technology addresses is structural. Blockchain wallets are identified by character strings rather than names, and crypto transfers cross jurisdictions at speeds that outpace traditional regulatory responses. As trading volumes have grown, so have the opportunities to layer and obscure the origins of funds — a challenge that conventional banking surveillance was never designed to meet.
The system was developed by researchers at the People's Public Security University of China, an institution tied to the Ministry of Public Security, whose findings appeared in May. The AI was trained on historical laundering schemes, enabling it to recognize similar behavioral patterns in new transactions. Its memory module builds relational maps of wallets over time, while the language model draws on contextual sources — news, regulatory filings, known criminal networks — to assess whether a transaction fits the profile of illicit movement.
The enforcement backdrop makes the technology's arrival timely. China's Supreme People's Procuratorate reported in March that prosecutors had indicted 3,259 people in 2025 alone for money laundering tied to virtual currencies and underground banking — a single year's figure that illustrates the sustained scale of the problem. Traditional crypto crime investigation is slow and expertise-intensive; an AI that reliably narrows the field could meaningfully accelerate the work of overstretched agencies.
The development carries implications well beyond China's borders. As cryptocurrency becomes more embedded in legitimate finance, regulators everywhere face the same dilemma — how to preserve the technology's utility while closing off its criminal applications. China's investment in algorithmic detection suggests a template others will follow, opening a new chapter in the technological contest between financial law enforcement and those who seek to exploit its gaps.
China's police forces have begun deploying an artificial intelligence system designed to catch bitcoin laundering, achieving what researchers claim is a 90 percent success rate in identifying illicit transactions. The framework combines a memory module with a large language model to track the movement of cryptocurrency across borders, cutting through the anonymity that makes digital currencies attractive to criminals in the first place.
The technology addresses a fundamental problem: as cryptocurrency trading has exploded in volume and complexity, so too have the opportunities for moving dirty money across jurisdictions without triggering traditional banking alerts. The pseudonymous nature of blockchain transactions—where wallets are identified by long strings of characters rather than names—creates a natural cover for those seeking to hide the origins of funds. The cross-border speed of crypto transfers compounds the challenge, allowing money to move faster than regulators can typically respond.
Researchers from the People's Public Security University of China, an institution affiliated with the Ministry of Public Security, published their findings in May. One of the study's authors, a researcher focused on criminal investigation and cybersecurity, described the AI system as offering "an innovative technological pathway for regulatory authorities to combat illicit cryptocurrency transactions and economic crimes." The framing reflects how seriously Beijing views the problem—not merely as a technical challenge, but as a matter of national economic security.
The urgency is evident in recent enforcement numbers. In March, China's Supreme People's Procuratorate announced that prosecutors had indicted 3,259 people during 2025 alone for money laundering connected to virtual currencies and underground banking operations. That single year's figure underscores the scale of the problem authorities believe they are facing. The crackdown has been sustained and systematic, with cryptocurrency-related financial crime remaining a priority across multiple government agencies.
The AI system works by analyzing patterns in transaction data that human investigators might miss or take far longer to identify. By feeding the algorithm historical examples of known laundering schemes, researchers trained it to recognize similar structures and behaviors in new transactions. The memory module allows the system to track relationships between wallets and entities over time, building a map of suspicious activity. The large language model component can process contextual information—news reports, regulatory filings, known criminal networks—to assess whether a particular transaction fits the profile of illicit movement.
What makes the 90 percent accuracy claim significant is that it suggests the technology could become a practical tool for overwhelmed enforcement agencies. Traditional investigation of cryptocurrency crimes is labor-intensive and requires specialized expertise. An AI system that correctly identifies nine out of ten suspicious transactions could dramatically accelerate the investigative process, allowing human analysts to focus their limited resources on the cases most likely to yield results.
The development also signals how enforcement strategies are evolving globally. As cryptocurrencies become more integrated into legitimate finance, regulators everywhere face the same dilemma: how to preserve the technology's benefits while preventing its use for criminal purposes. China's investment in AI-driven detection suggests other countries will likely follow, creating a new frontier in the technological arms race between law enforcement and those seeking to exploit financial system vulnerabilities.
Citazioni salienti
Provides an innovative technological pathway for regulatory authorities to combat illicit cryptocurrency transactions and economic crimes— Researcher specializing in criminal investigation and cybersecurity, People's Public Security University of China
Vuoi la storia completa?
Leggi l'originale su South China Morning Post ↗
Technology