Deep Threat Analysis: Going Beyond the Obvious to Improve Border Security

Print Friendly, PDF & Email

In this special guest feature, John Kendall, Director of the Border and National Security Program for the global public sector practice at Unisys, discusses how border security experts around the world can improve their threat analysis with new AI and machine learning tools. Based in Canberra Australia, John has overall responsibility for Unisys border and national security initiatives around the globe. This includes R&D, sales, consulting and delivery activities. With a 30-year career with Unisys, John has worked with Public Sector clients in the USA, Europe, Africa, Asia, Latin America and the South Pacific. 

To prevent threats, one must predict threats. Adversaries are constantly changing strategies and tactics, and many of today’s border protection approaches are limited to watch lists and known patterns that fall short.  A new approach, known as “deep threat analysis,” uses predictive analytics to examine the complex web of relationships and actions surrounding a traveler or shipment crossing a border. This yields better insights for thwarting bad actors.

To determine the risk associated with a traveler or shipment, deep threat analysis pulls data not only from airline and law enforcement records, but also from historical travel records or data regarding associates.

Deep threat analysis uses multiple techniques to provide a more insightful and comprehensive view.  For example:

  1. Analysis of complex relationships. Correlates and connects travelers and circumstances, such as others traveling on the same plane or to the same destination.
  2. Sophisticated engineered rules. Compares travel details against previous illicit activities.  For instance, DMV data can show if a person owns a car reportedly used during a crime.
  3. Statistical analysis or predictive analysis. Automatically detects and flags anomalies which may not be obvious to the human eye but which could indicate a potential threat.
  4. Continuous feedback or machine learning. Automatically updates threat assessment algorithms with the results from previous assessments – for improved accuracy with every assessment.

Putting Deep Threat Analysis into Action

The International Labor Organization estimates there are nearly 25 million victims of human trafficking globally. People who are trapped in this world are often from vulnerable populations – migrants and children – and that makes it all too easy for them to fall through the cracks.

How might deep threat analysis use data to find that one tiny thread and follow it to the end to track someone participating in human trafficking? Consider the following hypothetical (but all too common) scenario:

Sid is a registered sex offender in the U.K. who is banned from traveling overseas. Sid employs the dark web to “procure” a Ukrainian minor. A Russian syndicate arranges for a cohort (Noel) to deliver an unsuspecting minor (Sandra) to Sid.  Sandra believes Noel is taking her on holiday. The U.K. authorities have only one chance to detect and prevent the trafficking of Sandra — when she attempts to enter the U.K.

Traditional targeting solutions would assess Sandra and Noel as “no risk” because neither is listed as a person of interest on any watch list, and their passports and visa paperwork are in order. But deep threat analysis tells a very different story.

The deep threat analysis system flags an anomaly: Noel has traveled to the U.K. several times previously, each time with a different minor. Furthermore, using sophisticated engineered rules, the system would show that the airline tickets used by Sandra and Noel were purchased using a credit card associated with a suspected child trafficker.

This triggers an alert to a U.K. border security analyst to investigate Noel’s social media presence, which reveals that Noel is connected with others who are connected to a crime syndicate. An analysis of dark web chatter suggests that syndicate is currently planning to send a child to the U.K.

Based on this evidence, authorities in Kiev question Noel and Sandra before they board their flight and Sandra is taken to safety. Meanwhile, the system notes all these actions in real-time and flags another person of interest who is connected to a similar human trafficking ring in another country.  This creates a continuous machine learning feedback loop, whereby the system becomes smarter and more precise over time.

This is just one example of how deep threat analysis empowers authorities to predict threats with greater accuracy to improve border security methods and prevent crime.


Sign up for the free insideBIGDATA newsletter.

Speak Your Mind