Scientists developed a system that can detect money laundering transactions on the Bitcoin blockchain. The AI-based tool identifies not only individual transactions but also captures patterns of illicit operations involving BTC.

AI Model Detects Money Laundering Schemes on Bitcoin Network

Scientists from MIT-IBM Watson AI Lab and Elliptic conducted extensive experiments using artificial intelligence to detect money laundering operations on the Bitcoin blockchain. According to the research results, they created a specialized tool capable of identifying funds flowing into crypto exchange accounts from suspicious sources.

The main goal of the research was to explore the capabilities of AI in improving the mechanism for detecting financial crime schemes on blockchain networks. During the experiment, the machine learning model analyzed over 200,000 BTC transactions, marking those associated with illegal operations, such as ransomware schemes or the dark web. As a result, a transaction identification system was developed, which was later applied to analyze nearly 200 million BTC transactions. The AI tool learned to identify and mark entire chains of BTC transactions involved in money laundering. This system allows for the detection not only of individual crypto wallet addresses associated with illegal activities but also the identification and documentation of specific multi-level money laundering schemes.

To test the tool’s effectiveness, researchers tested it in analyzing BTC transactions on a crypto exchange, the name of which wasn’t disclosed. The model identified 52 cases of money laundering by exchange users through transactions with Bitcoin. Exchange representatives confirmed that 14 of the identified transactions were indeed associated with addresses flagged as suspicious. Interestingly, the exchange identifies and flags users based on off-chain data. The developed AI model can detect cases of money laundering that couldn’t be identified using traditional on-chain analysis methods.

The project authors also claim that the developed machine learning model can be used to identify previously unknown wallets associated with criminal activity. Since the AI tool can identify entire schemes, all addresses participating in this scheme will be automatically marked as belonging to criminals. This approach allowed researchers to identify several wallets used in Ponzi schemes and the dark web that were previously unknown.

Elliptic plans to use this AI model to improve its products and encourages the community to use the tool for data exchange to develop methods to combat crypto crime.

Despite a decrease in the volume of illegal crypto transactions in 2023, assets worth over $24.2 billion were involved in them. Moreover, more than 60% of illegal crypto transactions were carried out using stablecoins.

Author: Ana Bustos García
#AI #Blockchain #Hacking #News