Imagine a computer program that can automatically detect an onslaught of vitriol on social media aimed at U.S. military intervention in a foreign country, and then flood social media websites with information to counter this rising wave of discontent. That’s essentially the idea behind a DARPA project, started in 2011, called “Social Media in Strategic Communications,” or SMISC, which will “develop a new science of social networks built on an emerging technology base.” With social media websites like Twitter emerging as a force for social change, it’s not surprising that the Pentagon would like to come up with ways to track, and even influence, the world of social media, particularly in cases when social media is being used for deception, or to foster anti-American sentiment. But unlike conventional propaganda, produced by a few individuals or governments, social media might include a cast of thousands, and the number of messages in the tens of thousands. The goal of the project is to develop technology that can automatically detect “deceptive messaging and misinformation” and the come up with ways to counter that information, prompting Wired to call it a “social media propaganda machine.”
In 2008, DARPA held a competition, called the Network Challenge, offering $40,000 to the first team that could locate the position of 10 large red balloons released at fixed locations around the continental United States. The goal of the contest, according to the agency, was to “explore the roles the Internet and social networking play in the timely communication, wide-area team building, and urgent mobilization required to solve broad-scope, time-critical problems.” In other words, how can a social network be used collect highly accurate information as quickly as possible? The key to success, DARPA surmised, would be leveraging the immense power of the Internet and social networks to collect information about the balloons’ locations. The winning team, based at the Massachusetts Institute of Technology, won by offering what it called a “recursive incentive mechanism,” which involved using a sliding pay scale that financially rewarded not just the people who spotted balloons, but also their connections to those people who spotted balloons, creating an incentive to maximize growth of the network. The network challenge reflects DARPA’s move into social networking science, while also drawing on its prior experience in using competitions (such as robotic car races) with cash prizes to foster innovation and draw interest from nontraditional sources.
The video below shows how a team from Massachusetts Institute of Technology used social connections to locate the balloons.
While many social network analysis tools are still in development, DARPA has worked to move a program, called Nexus 7, to the battlefield. Nexus 7 was first described publicly by Wired in 2011 as “a classified and controversial intelligence program” being used to help track terrorists and insurgents in Afghanistan. In congressional testimony, DARPA Director Regina Dugan eluded to the program, calling it “a forward operating cell,” which “brought state-of-the-art advances in large-scale computational and visualization techniques to sparse database analysis, and created an in-theater data processing, exploitation, and dissemination cell complete with a full flight team.” In official budget documents DARPA has described Nexus 7 is more subdued terms as a tool for “automated interpretation, quantitative analysis, and visualization of social networks.” The program would be used to “develop and apply emerging methods for edge finding and cluster analysis to detect, characterize, and predict the dynamics of social networks,” the official description says. The only hint is of its operational relevance was an acknowledgement that social network analysis could offer a way to help track “terrorist cells, insurgent groups, and other stateless actors.” While DARPA has not provided many details on the program, Dugan, the agency’s director, has publicly lauded Nexus 7’s contributions to the war in Afghanistan.
While the science of studying social networks is not new, the rapid growth of the Internet, social media, mobile communications has produced an abundance of new data on how such networks work. DARPA in 2011 started a new research called Graph-theoretic Research in Algorithms and the PHenomenology of Social networks, or GRAPHS, designed to establish “foundational mathematics” for analyzing social networks with the goal of being to predict and influence the actions of such networks, such as terrorist networks “Recent successes in apprehending high-value individuals demonstrate that elementary methods of analysis can have a large operational impact when applied in this setting,” DARPA says. The goal of the program is to combine fields such as mathematics, physics and computer science to develop algorithms that can be used to predict these social networks.
DARPA’s Scalable Social Network Analysis (SSNA) project was a subject of the agency’s larger Total Information Awareness program, created in 2002 to help come up with data-mining technologies to fight terrorism. “The SSNA algorithm program will help distinguish potential terrorist cells based on their patterns of interactions from legitimate groups of people and identify when a terrorist group plans to execute an attack,” according to one description of the program provided by the Defense Department’s Inspector General. Since the Total Information Awareness program was canceled in 2003 amid public and congressional concerns about its privacy implications, it’s unclear how far the SSNA project actually went.