CMU Developing Online Tool to Identify and Track Down Sex Traffickers

Jan 19, 2015

Human trafficking is the second largest criminal enterprise in the world, second only to drug trafficking, according the FBI.

Sex trafficking is the most common form of human trafficking, and in an effort to try and find and identify those involved in the crime, researchers at CMU are developing online tools that go after a major vulnerability for sex traffickers — the need to advertise.

“At the end of the day, they have a service they’re trying to sell, and that means that they need to advertise it to their prospective clients,” said Jeff Schneider, a research professor at CMU and the project’s principal investigator.

But it’s not as though those advertisements can be obvious or straightforward. So the online tool being developed helps investigators follow up on small tips.

“Maybe it’s a phone number or something else, then have the tool retrieve any relevant information it can find related to that tip so the investigator can then follow up,” said Schneider.

The project builds on the work of researchers from the Auton Lab, which developed a program called Traffic Jam. More than 200 law enforcement officers have been trained in its use. But, Schneider says, better algorithms can be developed.

“To find patterns in the data, to look for not only just connections that are directly related to some tip, but also to find more subtle connections that wouldn’t be so obvious from a traditional search engine point of view,” he said.

It’s hard to track down and prosecute sex traffickers, and often when law enforcement catches on to something they can use to crack down, the criminals will change behaviors leaving law enforcement to start from scratch. So, the online tool is meant to be an ever-developing technology.

“What’s needed is not only a one-off effort to figure out some magic keywords, but really a learning algorithm that can constantly monitor content and try to understand how the keywords are evolving over time and what the new ones may be,” said Schneider.

The project is being funded through a contract from the Defense Advanced Research Projects Agency ‘s (DARPA) Memex program, a three-year research initiative.