“If you can track the money, you can track the people.”
—Anjuli Bedi, Associate Director of Psychometrics and Predictive Analytics
That’s why John McGrath built the Traffik Analysis Hub: to help financial organizations identify where human traffickers are benefitting from trafficking, and ultimately disrupt human trafficking networks at their source.
And he’s doing it through data: first gathering data at scale, then using AI to refine and enrich that data.
“So that we can identify global hot spots,” he said, “and identify where trafficking routes happen, what the transit points are, what types of trafficking are prevalent in different time periods and different locations.”
The global estimates of human trafficking vary precisely because of the nature of the crime: the victims disappear, uncounted and unreported. It’s estimated that around 40 million people are in circumstances of human trafficking in a given year. The global revenue, meanwhile, is estimated at around $150 billion a year—making this the third-largest and fastest-growing criminal activity in in the world.
McGrath, an IBM senior solution architect, began building the Hub in 2016. He was facilitating a workshop in London for STOP THE TRAFFIK, a global non-governmental organization focused on human trafficking prevention. Attendees included law enforcement agencies, NGOs and, importantly, financial institutions. They all identified a need to share data.
McGrath found some relevant data sets online. He built prototypes, and then worked with STT to build some more: a map interface with hot spots, heat maps, markers. He shared the prototypes in a follow-up workshop.
“We gave them a vision for how this system could be brought together,” he said.
One of STT’s core goals is raising awareness of human trafficking through providing a clearer look into what’s happening on the ground globally.
The work of McGrath and his team has transformed the way STT works to achieve that goal.
“Everybody holds part of the story,” STT CEO Ruth Dearnley said. “Whether the person in the village; on the street; in the bank; or the person trying to prosecute and arrest. We needed to bring those stories together, to weave them into a tapestry that would show us a story bigger than any one person held. IBM gave us that possibility.”
A crucial component of that story is money. It’s estimated that one percent of criminal proceeds from trafficking are confiscated or disrupted, according to financial crime expert Geraldine Lawlor.
“That is an extraordinarily high profit margin and why it is so lucrative to those involved,” she said.
McGrath and his team are working with financial institutions for that reason: if you follow the money, you can also follow the flow of humans being trafficked. Money laundering is a prevalent part of human trafficking, so they’re identifying patterns of anomalies in transactions related to laundering.
“We’re starting to identify patterns developing in specific locations for specific types of transactions, and for specific types of trafficking,” he said.
That’s something financial institutions have never had access to before: pooled data from multiple sources—NGOs, publicly available news via AI and other peer financial institutions. That gives them a view beyond their own internal horizon. Which, in turn, allows them to better focus the microscope.
In the Hub, each authenticated participant gets access to a map interface, news explorer, and analysis register.
The map is important, because it generates hot spots based on different representations of data. That data has come into the Hub through various ways; including NGOs like STT sharing their own manually curated data with McGrath. AI engines, including IBM Watson, are another source.
Consumers of the data include law enforcement agencies: “They pull the data offline and do forensic analysis on it,” McGrath said.
At STT’s office last month, McGrath walked me through the Hub. On the map interface, he clicked through various menu items.
“These are clustered markers,” he explained. “They give us a count, intensity color and time period. We’re looking at the data we have from Watson, Google Earth and the NGO community.”
McGrath pointed out clusters of intensity in the Indian subcontinent and the UK.
“If we were to zoom in and break these clusters apart, we can see some of the raw data behind them,” he said.
The Hub doesn’t store any personal or sensitive information. The data is stored in the IBM Cloud, and because it’s non-specific, the jurisdiction in which the data is located isn’t an issue.
McGrath clicked on another part of the map.
“An additional important feature here is the location type,” he said. “In this case it’s a transit point.”
McGrath zoomed into a transit point related to an event that occurred in Romania.
“The source of the incident is Romania, and it’s connecting to both Dublin and Northern Ireland,” he said. “When you look at the detail behind this particular incident, it’s people being trafficked into the agricultural business in Northern Ireland via Dublin.”
On the screen were lavender dots, crisscrossed lines. They connected across cities, countries, continents. Humans, being bought and sold and moved.
NGOs like STOP THE TRAFFIK have never had this kind of information before: stories being woven together into a more complete picture.
“This system can tell you where the transits and the sources are,” McGrath said. “You can find out which other partner NGOs have more detail, or more data and specialty in the sources. So you can form collaborations.”
McGrath is working on a prototype for financial institution data so that banks too can collaborate and share data. The prototype is based on red-flag indicator data for transactions that have exceeded the Thomson Reuters Foundation rules.
The non-attributed data doesn’t contain specifics around entities or persons. It does tell when the transaction occurred, the rule or trigger it fell afoul of, the source and destination cities.
McGrath uses this kind of data to do a proximity correlation.
“We can overlay for a particular type of trafficking and a particular period in time,” he said. “The financial anomaly data can be overlaid with the heat map of the incidents from the NGO community.”
The Hub highlights particular patterns in the financial transactions. McGrath demonstrated a transaction pattern flowing from Dublin through to Lagos. It’s highlighted because of physical proximity to an old hotspot for trafficking. With this capability, the Hub is providing financial institutions with a notification that, in the haystack of potential transactions, this one may be where the needle is located.
“Where did they go? How? Who gave the money? Who took their passport? What borders did they cross?” Dearnley said. “With IBM we’ve begun to build that data, identify patterns, identify hot spots.”
The next milestone McGrath is working into the Hub is projection: looking at patterns in the data that exist historically, and seeing if those patterns are being repeated in real time. One day, perhaps, the Hub will be able project what may happen in the future.
“The transactional information is real time or near-real time,” McGrath said. “If we can identify past patterns when known events occurred, we can possibly match those patterns to what we’re seeing developing right now. Which may give an indication if something’s occurring right now that we’re not aware of yet.”