Bitcoin & AI: Trap for Sex Traffickers

Source: Cristian Newman, Unsplash

Rebecca Portnoff from Berkley University created an algorithm that can track sex traffickers with the help of Bitcoin.

Many prostitutes do not offer their services voluntarily. They are often abducted, sold and forced into prostitution. This range of services is shifting more and more to the Internet in order to offer the prostitutes convenient and secure access to human goods. Bitcoin and science can help distinguish sex trafficking from voluntary services and uncover the people behind it.

Step 1: the offer

There are a lot of websites with sex offers. One of the largest in the world is the classifieds site Backpage.com. Previously, detectives had to manually read thousands of ads a day to find out where sex trafficking might be involved. This is very time-consuming and, of course, very stressful for the psyche of these readers. But what is the evidence for sex trafficking? If one and the same author creates many ads with many different sex offers, the probability is high that among them are involuntary sex services.

Step 2: Artificial intelligence

Rebecca Portnoff from the University of California, Berkeley, and her team have developed an algorithm using machine learning. The software is fed text patterns that can indicate sex slavery. Then the software finds suitable offers. With each search, the algorithm refines itself. 
“Our technology finds connections between the ads,” explains Portnoff. “Is the pimp behind the Backpage ad also Craigslist’s pimp? Is he the same man who gets Bitcoin for enslaved sex workers? These questions can only be answered with advanced technology like our algorithm.” The first part of Portnoff’s software is based on Stylometry — the analysis of text parts to determine whether two texts come from the same author using language usage patterns.
“There are hundreds of thousands of these ads placed each year,” adds Portnoff’s colleague Damon McCoy. “Any technique that brings commonalities between ads to light and potentially illuminates their owners is a big push for those working to contain exploitation.”

Step 3: Enter Bitcoin

Since Backpage is the world’s most widely used portal for sex services, Visa, Mastercard, and others deny the operator the opportunity to offer online payment by credit card. Therefore, the ads are most likely to be paid with Bitcoin.

Step 4: No anonymity whatsoever

The Bitcoin transactions are stored in the public blockchain. This allows the Portnoff team to calculate which Bitcoin addresses are displaying many different ads. This is a strong indication that there is a larger organization behind it and that the likelihood of sex trafficking is increasing. But how is this possible? You can see in the blockchain that the address A sends Bitcoin to the address B, but you don’t know who A and B are.

Source: Kristina Flour, Unsplash

Step 5: Sherlock Holmes for mathematicians

The way Portnoff does it is worthy of Sherlock Holmes. Usually, a Bitcoin transaction goes like this: Address A sends address B Bitcoin. Address B registers this and waits for the transaction to be validated by the entire network. This takes between 10 minutes and several hours. Only then is the transaction complete. The webpage Backpage.com does not want to wait that long. Because the customers fill out the profile, send Bitcoin and should get it online immediately. Backpage thus risks that the transaction is not validated. They don’t care, so the transaction is completed immediately. In the blockchain, the transaction is timestamped to the millisecond and on the public profile of Backpage.com. Portnoff now takes the Backpage timestamp and compares it with all Bitcoin transactions in the blockchain. In this way, it finds related transactions. It then groups Bitcoin addresses and can find out if many ads are paid for from one address.
This is a shortened representation. In Portnoff’s scientific paper, it spreads the process out over ten closely described pages.

Step 5: Bring in the Feds

The data collected by Rebecca Portnoff can now be easily linked. Were the Bitcoin addresses used to pay for other things that might lead to telephone numbers or email addresses? Are the Bitcoin purchased in exchanges with a credit card and can these data be viewed by court order? Portnoff says that NGOs and authorities are either already very interested in their program, or are already using it.
Incidentally, Backpage.com’s CEO Carl Ferrer was arrested in October 2016 for allowing forced prostitution. The trial is still ongoing at the time of writing even though the allegation has been reduced to money laundering.

Read another tale of Bitcoin & Crime:

https://medium.com/@bitcoinblase/the-silk-road-a-real-thriller-and-the-truth-about-the-anonymity-of-bitcoin-198b519ca397

Sources

17 https://people.eecs.berkeley.edu/~rsportnoff/

18 http://news.berkeley.edu/2017/08/16/in-a-step-toward-fighting-human-trafficking-sex-ads-are-linked-to-bitcoin-data/

19 https://people.eecs.berkeley.edu/~rsportnoff/resources/KDD-Final-Portnoff.pdf

20 http://fortune.com/2017/08/25/bitcoin-ai-sex-trafficking/

21 https://cointelegraph.com/news/backpagecom-ceo-carl-ferrer-arrested-charged-with-being-a-pimp

22 http://www.washingtontimes.com/news/2017/jan/10/carl-ferrer-backpage-ceo-mum-screening-policies/

23 https://reason.com/blog/2017/08/31/california-drops-kamala-harris-pimping

This story is published in The Startup, Medium’s largest entrepreneurship publication followed by +397,714 people.

Subscribe to receive our top stories here.

Leave a Comment