Human trafficking is one of the most heinous crimes that one group of humans can engage in over another, with the practice taking the form of modern-day slavery.
The extent of the human trafficking problem in the U.S. is considerable . standing at just under 12,000 cases per year. This is dwarfed, however, by the international scale of the crime, with an estimated 25 million people trafficked globally. This produces $150 billon for the crime gangs involved, according to World 101 data.
To help in the global coordinated action underway to counteract human trafficking, researchers at MIT’s Lincoln Laboratory have been working on technological solutions. This part of efforts to harness technology in order to help identify, investigate, and prosecute trafficking cases.
Technology is needed not least because the criminal elements involved have themselves made effective use of digital solutions to help construct and operate their criminal networks. This includes operating illicit sex digital marketplaces to using the Internet for complex money laundering networks.
The MIT efforts are being led by Dr. Matthew Daggett. This includes harnessing article intelligence to help make sense of the complex web of digital data that arises from human traffic investigations as well as to decipher the attempts by criminals to hide their trails.
To help to develop suitable systems and to define the scope, Daggett established a forum called the Counter–Human Trafficking Technology Workshop. The event drew together 200 personnel from 85 U.S. federal and state agencies, together with activists from non-governmental organizations and academics.
The forum considered what types of technology are available today and as a result of identified gaps, the types of technology needed to take the battle forwards.
One of the most promising technologies is based on AI, drawing on machine learning algorithms. Success has been shown through the analysis of Internet adverts offering sexual services. The analysis helps to reveal and then pinpoint together whether a given advert is associated with human trafficking and how individual adverts piece together. This latter approach helps to map out the criminal network. Challenges arise as the same sex service is often advertised across multiple sites, making its origin and association with other adverts difficult to discern.
The way around this is where the machine learning algorithms can assess the language used in thousands of adverts, looking for similarities, certain nuanced phrases and so on. The pattern analysis helps build an idea of relatedness. Similar strides are also being made with image analysis.
The aim of this approach is help unravel large-scale, organized trafficking networks. With this it is seen as important that the whole network is mapped and then a blow struck, for simply hitting a node within the network sees the criminal chain fairly easily recover.
A strength is with the ‘learning’ aspect of the AI. Success in identifying a criminal gang enables the algorithm to fine-tune and this enables the program to rate the likelihood that an advert is linked to human trafficking, which helps law enforcement to better direct their efforts.
