Machine Learning to Combat Human (Sex) Trafficking

tl;dr
Algorithms identify bitcoin wallets controlled by sex traffickers:

http://www.prnewswire.com/news-releases/follow-the-bitcoin-to-find-victims-of-human-trafficking-300503503.html

Image credit Bakary Seckan - image found on Wiki Commons

My summary below is in my own words with my own conclusion. Copy pasta is a dish best served... never :)

The researchers used two different algorithms to identify potential ringleaders of trafficking rings. The first algorithm looks at writing style and words used (stylometry) to identify when a single author writes multiple different online sex ads (posing as different people) - the researchers tested the algorithm by uploading some of their own control ads and claim an 89% true positive success rate in identifying authors of multiple posts.

The second algorithm compares the time stamps of when these ads showed up online with time stamps of bitcoin transactions, looking for correlations that would link specific bitcoin wallet addresses with the individuals writing and posting these ads.

While there are a lot of really great things about decentralization and blockchain technology, there is no denying that the rise of Bitcoin et al has made life a lot easier on organized crime. Human sex trafficking, often of minors, is a very real problem that is a lot more prevalent than you might expect if you have not researched the topic. Taking the machine learning, algorithmic approach here is a great first step in providing law enforcement agencies with the tools needed to effectively combat this.

Cheers - Carl

H2
H3
H4
Upload from PC
Video gallery
3 columns
2 columns
1 column
7 Comments