GLOBAL CHALLENGES

Army Dragnet Targets Terrorists on Social Nets

R. Colin Johnson | Date: 05-19-10 | Comments

At Rensselaer Polytechnic Institute using IBM's Blue Gene supercomputer, efforts to model how terrorists recruit ordinary citizens over the social networks have hopes of unveiling the steps by which ordinary citizens are radicalized online, so that these trends can be detected and countered before tragedy ensues.

The World Wide Web permits anyone with a browser to access any computer system on the globe anonymously. As a result, citizens have quick, fear-free access to information, but at the cost of leading users to anonymous "sessions" that could be conducted by recruiters for terrorist organizations.

The IBM Blue Gene supercomputer at RPI will be enlisted by the U.S. Army to identify, track and counter terrorists� online recruiters.

As a result, online communities have become a free-for-all where plots are hatched, organized and executed all under the watchful eye of traditional security organizations, but with no easy way for law enforcement officials to track down and arrest potential terrorists before, during or after the fact. Now, the U.S. Army Research Laboratory aims to improve the situation by funding Rensselaer Polytechnic Institute in a 10-year, $35 million probe at its newly established Social Cognitive Networks Academic Research Center (SCNARC).

SCNARC is enlisting engineers and scientists to analyze the "signatures" of global terror networks in search of inventive methods of detecting and tracking them down within the constraints of the anonymity of the Internet. To test out scenarios, SCNARC is creating a virtual network on IBM's Blue Gene supercomputer that utilizes thousands of software agents to simulate "adversarial users" interacting with ordinary people. Its goal is to understand how terrorists succeed in radicalizing normal citizens via global networks, and to find ways of detecting these efforts before they turn impressionable youngsters into suicide bombers.

The first phase of the program has already been allotted $16.75 million for 2010 to 2015, and if all goes well, it will add $18.75 million for the last five years (2015 to 2020). The contract calls for RPI to unveil how enough trust can be built among online users to radicalize their thoughts and actions. To quantify its search, electrodes will be attached to real users interacting with virtual terrorists online to observe the physical effects of radicalizing cognitive interactions. The virtual network of software agents will be programmed to express the cognitive states that RPI researchers believe contribute to radicalizing previously peaceful citizens. If RPI can thus identify the signature of how this trust is built among strangers, then it hopes to enable the early detection of such attempts by security personnel monitoring online transactions.

Besides trust building, SCNARC will also model dynamic processes (how interactions influence people), organizational networks (how hierarchies of influence arise), game theory (ways of countering adversarial networks) and the applicable principles of human cognition that facilitate understanding.

Partners with RPI on the project include IBM, Northeastern University, the City University of New York, as well as participants from Harvard University, Massachusetts Institute of Technology, New York University, Northwestern University, the University of Notre Dame, the University of Maryland and Indiana University.