Sex trafficking and human slavery are certainly nothing new, but the internet has created a lawless space for predators to buy and sell people. Today, more than 150,000 escort ads are posted in the US every day, many of them for children. The human trafficking industry enslaves an estimated 27 million people worldwide.
Now, an organization is turning to the very features of the internet that make trafficking so widespread to combat it. Thorn partners with technology companies like Google, Pinterest, Facebook and others to help identify and rescue children, and possibly catch predators.
Last month, a video of Thorn founder Ashton Kutcher went viral when he gave very personal testimony to the US Senate on the tragedy of child trafficking. Earlier this year, Sen. Bob Corker of Tennessee sponsored a bill that eventually became a law to direct $50 million in federal money annually to combat sex trafficking and human slavery in the US.
Following Kutcher’s Senate speech in February, the White House — led by Ivanka Trump — invited a number of anti-trafficking groups in to discuss what can be done. Policy proposals may follow.
Traditionally, law enforcement efforts to combat trafficking have been insufficient, given the fluid nature, anonymity and enormous reach of the internet. Thorn’s approach is interesting because it creates tech tools specifically geared to helping the authorities.
In 2011, law enforcement officials in the US turned 22 million images and videos of child abuse over to the National Center for Missing and Exploited Children to identify the victims, according to the center's US Sentencing Commission testimony cited by Nicholas Kristof in The New York Times. In the US alone, 9 million computer IP addresses were tracked sharing child pornography files. The sheer volume of victims and perpetrators is overwhelming for law enforcement.
Child abuse images are often traded on peer-to-peer networks or inside password-protected chat rooms. Most buying and selling of sex occurs online, on listings sites like Backpage or Craigslist — where escort ads are posted, and customers text in to make contact. That’s not illegal. The challenge for law enforcement is that many of those featured in the ads may be underage girls who are trafficked — mixed in with women posting ads voluntarily.
“These transactions don’t happen in the open,” says Julie Cordua, CEO of Thorn. “I can go on to backpage.com and put my couch up for sale and if I click the next tab I can buy a 14-year-old for sex, and this is not illegal, because an escort is not illegal. Every day, hundreds of thousands of ads are posted, and some are children, but how do we know? It’s like trying to find a needle in a haystack.”
Innovations in tech and data science can make that process more efficient.
Thorn is using machine learning, in which computers learn what advertisements represent a child, and create an algorithm to predict what other ads are more likely to be associated with a child. That, they hope, can reduce the thousands of images of children in circulation.
They’re also using facial recognition software that recognizes signs of aging and can identify children from photos. They are working with the National Center for Missing and Exploited Children's database of missing children that could be matched to images from online ads. This helps detectives in tracking down and identifying children.
There is also a new texting application called Befreetexting designed to reach trafficked children who have sporadic access to cellphones. “These kids are held kind of captive, so they can’t pick up a phone, but they can text, and so we can create a text hotline,” Cordua says.
By sorting through the images to identify which are children, law enforcement can hone in on victims. The goal is to get to children as quickly as possible. That's just one side of the problem. Creating technology to go after the predators, either the customer or the pimp, has proven more difficult.
Christina Asquith is director of Across Women’s Lives, and founder of The Fuller Project for International Reporting.