Critics of Bitcoin have long argued that cryptocurrencies have created a haven for criminals. Because of blockchain’s distributed nature and its convenience for anonymous transactions, it’s challenging for law enforcers to tell when a transaction is related to criminal activities such as money laundering or drug trafficking.
In 2017, for example, a University of Oxford study found that nearly half (44%) of Bitcoin transactions were associated with illegal activities. But since then, that proportion has dropped dramatically.
Last week, a dataset published by blockchain analytics firm Elliptic in collaboration with researchers at Massachusetts Institute of Technology’s IBM Watson AI Lab found that only 2% of Bitcoin transactions in 2019 were deemed “illicit.”
That finding was based on an analysis of over 200,000 Bitcoin node transactions worth about $6 billion. Among the remaining 98% of the published data, 21% was identified as lawful while 77% remained as unclassified data.
Using machine learning technology to identify patterns of illegal Bitcoin usage, this academic study provides a new tool for law enforcement to detect criminal activities on blockchain networks.
“A big problem with compliance, in general, is false positives. A big part of this research is minimizing the number of false positives. The key finding is that machine learning techniques are very effective at finding transactions that are illicit,” Elliptic co-founder Tom Robinson told reporters.
For example, Elliptic’s software is sometimes able to find patterns in Bitcoin transactions that are difficult to describe yet match with known illicit usages, based on data from the dark web and past criminal cases.
And yet, the absolute amount of Bitcoin spent on criminal activities is on the rise, despite its declining share of total Bitcoin transactions. Per a recent study by blockchain analytics firm Chainalysis, Bitcoin use in illegal online marketplaces—from drug trafficking to child porn—is set to hit a record high of $1 billion, partly thanks to Bitcoin’s rebounding value. About $500 million of Bitcoin has already been spent on the dark web in 2019 so far.