- A machine learning model created by Elliptic can spot money laundering trends.
- The MIT-IBM Watson AI Lab collaborated on the study, which is being used to improve the company’s offerings.
- The tool found 52 suspicious subgraphs in total without gaining access to any account data.
A machine learning model created by Elliptic can spot money laundering trends that conventional blockchain analytics would overlook.
With the availability of a dataset comprising more than 200 million transactions, the business has enabled the broader community to create alternative methods for identifying unlawful activities. The MIT-IBM Watson AI Lab collaborated on the study, which is being used to improve the company’s offerings.
Blockchain
A machine learning model was developed to identify “subgraphs,” or sequences of transactions that signify the laundering of Bitcoin. To find out if this novel strategy would effectively reveal attempts at money laundering within the company, a cryptocurrency exchange was brought in.
The tool found 52 suspicious subgraphs in total without gaining access to any account data. Subsequently, it became known that 14 wallets had been reported by the trading platform for possible illegal behavior. Typically, only one out of every 10,000 accounts has possible misconduct identified, indicating that the model operates exceptionally well.
Since the machine learning algorithm was able to discover wallets that were previously unknown, more investigation can be done to determine who is responsible.
Elliptic contends that while traditional finance’s “siloed” data makes such approaches less effective, cryptocurrencies are “far from being a haven for criminals” thanks to AI-based analytics and the transparency of blockchains.
Scammers, fraudsters, and hackers may find it much harder to target cryptocurrency if there were tools like these available to shorten the time it takes to identify illegitimate digital assets. This would help the industry overcome its negative reputation as a “Wild West” for these types of activities.
