Bitcoin and Artificial Intelligence Can Help Fight Sex Trafficking

Computer science researchers at the University of California, Berkeley have developed new tools to identify sex trafficking rings, making them easier for law enforcement to target and prosecute.

Those efforts have been stymied, according to the researchers’ report, by the vast quantity of ads for sex posted to websites like, only a portion of which may point to human trafficking or sex slavery. Screening thousands of new ads every day can also take a mental toll on human workers. The new technique, developed by a team including PhD candidate Rebecca S. Portnoff, combines two distinct approaches to solving that problem. First, the team created a machine-learning filter that finds stylistic similarities between ads for sex services posted to sites like That makes it easier to distinguish between women voluntarily engaging in sex work, and those being forced into it by criminal organizations posting multiple ads.

The second technique goes even deeper. Because Backpage is the most used portal for advertising sex services, credit card processors like Visa refused to service the site starting in 2015. That left Bitcoin as the preferred means of paying for ads. The Berkeley researchers took advantage of Bitcoin’s public blockchain — a record of all transactions — to identify payments for sex ads originating with the same Bitcoin user, again providing evidence of a larger organization and likelihood of trafficking.

Get Data Sheet, Fortune’s technology newsletter.

Those linkages could be used to subpoena further records from Bitcoin wallet services, or to expand on sting operations conducted using phone numbers or other contact information listed in the ads.

Bitcoin was initially hailed as an anonymous way of paying for services, making it hugely popular among users of Darknet markets like Silk Road and, later, AlphaBay. But law enforcement and cybersecurity researchers have shown that it’s not hard to track and trace Bitcoin transactions, and even to connect them to real-world names. That has helped bring down dark markets, while also spurring the development of less traceable cryptocurrencies like Monero and Zcash.

The Berkeley researchers said that, after encouraging tests, NGOs, companies, and law enforcement agencies are either already using, planning to use, or expressing strong interest in the new techniques.

This is part of Fortune’s new initiative, The Ledger, a trusted news source at the intersection of tech and finance. For more on The Ledger, click here.