Hamsa Bastani, assistant professor of operations, information, and decisions at Wharton, discusses key findings from her research on how deep web data and machine learning can help law enforcement agencies investigate and intervene in human trafficking.

In Wharton Social Impact’s “Research Spotlight” series, we highlight recent research by Wharton professors and doctoral students whose research focuses on the intersection of business and impact.

This month, we spoke with Hamsa Bastani, assistant professor of operations, information, and decisions at Wharton.  

Your study explores how deep web data and machine learning can uncover the risk for human trafficking. To begin, can you define what human trafficking is and isn’t?

Hamsa Bastani: As defined by the U.S. Government, human trafficking is “modern day slavery.” Human trafficking includes both sex trafficking and forced labor. In our research, we focus specifically on sex trafficking. Victims of sex trafficking are forced or deceived into engaging in sex acts for money.

How pervasive is human trafficking globally?

Bastani: According to the FBI, sex trafficking is the fastest-growing organized crime business and the third-largest criminal enterprise in the world. The International Labor Organization estimates there were 4.8 million sex trafficking victims in 2017 alone.

Why were you interested in researching this topic?

Bastani: Machine learning is increasingly giving us a lens into previously opaque problems by finding useful signals in large volumes of noisy data. My PhD student Pia Ramchandani and I are passionate about using these tools towards making a positive societal impact, which drew us to our collaboration with Tellfinder Alliance.

How do you gather the data, given the covert nature of human trafficking?

Bastani: We use data from leading commercial sex advertisement websites, obtained from our partner Tellfinder Alliance. We capture trafficking risk by linking deceptive non-sex recruitment offers (e.g., purportedly for modeling or massage) to commercial sex sales by the same entities. These are likely instances where traffickers are using deceptive advertising to lure vulnerable individuals into situations where they can be trafficked.

In a nutshell, what are your most important findings?

Bastani: Based on the above data, we construct the first global view of trafficking risk in commercial sex supply chains (i.e., connecting recruitment and sales). This allows us to infer likely recruitment-to-sales trafficking routes of criminal entities, deceptive approaches used to recruit victims, and regional variations in where recruitment versus sales occur. These insights can help law enforcement agencies operating near likely trafficking routes to better coordinate efforts; they can also help target local counter-trafficking policies and educational interventions to counter deceptive recruitment efforts used frequently in that region.

What surprised you most?

Bastani: While sex sales predominantly occur in large urban centers, we find evidence that recruitment is concentrated in suburban, economically constrained areas. These areas tend to have vulnerable populations with limited opportunities, making them attractive targets for traffickers to lure with deceptive recruitment offers. However, anti-trafficking efforts are largely focused on urban centers. Our results suggest that resource-constrained “recruitment hubs” may benefit from collaborations with (better-funded) counter-trafficking agencies in large cities.

What is the most important lesson or takeaway for law enforcement, policymakers, social workers, and/or other decision-makers? What are some ways these crimes can be prevented?

Bastani: First, inferring likely trafficking routes can help law enforcement agencies along those routes better coordinate efforts. The FBI reports that the most effective way to investigate human trafficking is through a “collaborative, multi-agency approach with our federal, state, local, and tribal partners.” Our analysis can help prioritize partnerships between impacted law enforcement jurisdictions.

Second, identifying region-specific exploitative behaviors can inform targeted local policies and interventions. Social policy plays an important role in preventing vulnerable victims from being trafficked, as well as rehabilitating victims after their trafficking experience. Our results provide large-scale insight into where and how victims are (often deceptively) recruited. Cities with significant rates of recruitment into trafficking may prefer to focus their resources on preventative measures and can furthermore tailor interventions towards the recruitment tactics frequently seen in their specific locale.

Posted: February 18, 2022

Read More Stories