The Benefits and Limitations of Crowdsourcing HUMINT
By Camelia M. Kantor, Ph.D.
Jan. 12, 2017
“But you are a geographer!”
“Exactly,” the geographer said. “But I am not an explorer. I haven’t a single explorer on my planet. It is not the geographer who goes out to count the towns, the rivers, the mountains, the seas, the oceans, and the deserts. The geographer is much too important to go loafing about. He does not leave his desk. But he receives the explorers in his study. He asks them questions, and he notes down what they recall of their travels. And if the recollections of any one among them seem interesting to him, the geographer orders an inquiry into that explorer’s moral character.” – from The Little Prince by Antoine de Saint-Exupéry
Human intelligence (HUMINT) is broadly defined as any information that can be gathered from human sources (CIA). Among all other intelligence collection disciplines, only HUMINT utilizes people as a source. Human traits have been analyzed and defined by many, but still puzzle us due to their uniqueness and complexity. Language, intelligence, morality, love of art and music, dreams, compassion toward other humans, animals, and nature—hence, “Everything is vague to a degree you do not realize till you have tried to make it precise.” And then, “it all comes back to logic in the strictest and most formal sense” to allow us to see that what we thought to be true is not necessarily what logic ends up finding because human sources are far from being precise as groups; their traits vary from individual to individual.
Gathering HUMINT is both a dangerous and challenging endeavor. Deep knowledge of the local population is of major importance in understanding adversaries’ capabilities. Still, HUMINT is being neglected or underrepresented in the U.S. intelligence collection efforts worldwide. And with advances in technology relying on Unmanned Aerial Vehicles (UAVs) and crowdsourcing to become ultimate data gathering platforms, sending intelligence officers to gather HUMINT might become a thing of the past.
The Unreliability of Crowdsourcing for HUMINT
Crowdsourcing is the practice of obtaining information (ideas, services, content) by soliciting free or low-pay contributions from the online community. The idea is to take the work usually performed in a traditional manner with full- or part-time employees and “outsource” it to the “crowd,” mostly volunteers who contribute their time and effort in support of a cause. As a business model, it allows companies to feel the crowd beyond its own employees’ input and collect faster and more diverse—maybe even more creative—data and information at a lower cost. Despite numerous successes registered recently by private companies and governments around the world, the process of harnessing the power of the crowd is not yet clearly outlined, leaving place for major debates regarding the ethics behind it. Having access to a large population with a variety of skills and capabilities for a fraction of the regular costs might sound great, but is it really the way to go in the Intelligence Community? There is already sufficient evidence that the community’s reliance on interconnected technology has made the U.S. more vulnerable to threats (see Symantec’s 2016 Internet Security Threat Report on the number of zero-day vulnerabilities). While the industry keeps learning from mistakes and tries to invent new ways to preserve adequate security levels, it seems that capable hackers are everywhere, including regions of the world that were late arrivals to the technological revolution. If the cybersecurity field did not yet find clear solutions to its issues, why would one consider harnessing the “crowds in the cloud” to solve sensitive matters of national or even international security?
Let’s take for example the Sub-Saharan African region and the country of Afghanistan. Despite economic challenges, Sub-Saharan Africa has 88% of the overall population (including 79% of its rural population) covered by a mobile cellular signal. Afghanistan has a rapidly expanding communications infrastructure with more than half of the population (54%) owning cell phones. Despite good or rapidly growing cellphone coverage, 60% of the Sub-Saharan African population is expected to lack internet access by the end of the decade based on current forecasts, while “Afghans in the most troubled, insurgent-held areas [still] live in information wastelands dominated by militant propaganda.” In this context, while crowdsourcing might seem an impossible method of HUMINT collection, with help on the way through “strategic” partnerships (Vodafone’s Internet of Things, Facebook’s and Eutelsat’s 2016 satellite launching and custom drone, or Google’s Project Loon), Sub-Saharan Africa might gain access to internet much sooner that initially predicted. Internet censored countries such as Afghanistan, Pakistan, and Iran might also reconsider their involvement with the cyber world.
So, what would this global internet revolution mean to the individuals living in areas targeted for HUMINT gathering? Will such initiatives transform conflict prone communities around the world into real-time contributors and volunteer consumers of data to tackle pressing challenges posed by political conflicts and/or climate change? In other words, to use a paraphrased 2013 headline from the Harvard Business Review: “Will the Crowd feel treated (not “used”) like an Innovation Partner and join in?” If everything goes as planned, a characteristic of an efficient society, the logical answer would be “Yes!” But in places where government effectiveness and regulatory quality, two aspects of governance found to matter in influencing technological efficiency, still present challenges, and where human capital is not yet robust, presenting numerous socio-economic and domestic political vulnerabilities, the logic would place ongoing crowdsourcing efforts under the refuted argument. We have very recent proof in the latest international events where three Iranian-American men and a Palestinian girl were sentenced to prison in Iran and Israel for their Facebook posts. Even if assuming society regulates itself to a certain extent, (there has been some evidence of structural reforms and privatization programs) will humans act like social entities and contribute their time, knowledge, and wisdom to the greater good?
Crowdsourcing for the Greater Good
“The greater good” has become a core theme in much of today’s public as well as scientific dialogue. A Howard University student told the Associated Press during a riot, “it’s a time for belief in the greater good of humanity.” The University of California, Berkeley is sponsoring and conducting ground-breaking research as part of its The Greater Good Science Center, exploring, among others, the roots of strong social bonds and altruistic behavior. Even USGIF prides itself for “working collectively toward the greater good.”
While these are great examples of how people and various organizations are making efforts to tackle together today’s world challenges as a greater, better, and more compassionate and altruistic community, on the other side we see a surge in pseudoscience propaganda messages tweeted, shared on Facebook, or presented in a growing number of movies urging people to stop vaccinating their children, to change their dietary habits based on unproven/unscientific claims of ultimate cures, to take on religious or creationist propaganda, and more—all in the name of the greater good. Faced with such an avalanche of information coming from experts, pretend experts, or simply ordinary people who seem passionate and genuine in their beliefs, the crowd is expected to decide what’s best for them. Considering many of them, especially the younger generations, are mainly connected to the outside world through this tumultuous, daily sharing of information, the power of the crowd over the crowd is at an apogee. Rapid volumes of shared information leave little time for in-depth research for truth as the crowd in cyberspace is no longer separated by educational attainment or wealth and less likely to have a close relationship with those sharing their “knowledge” through such platforms. This leaves people vulnerable to decision-making based solely on the “wisdom” shared by certain groups, further resulting in less desirable outcomes.
To give an example, despite the major progress registered by WHO’s Expanded Programme on Immunization since its inception in 1974, the CDC reported a surge in measles cases (a disease declared eliminated in the U.S.) from December 28, 2014, through January 21, 2015among Orange County, California, Disneyland Resort Theme Park visitors. Of the 52-outbreak associated cases 55% were unvaccinated and 31 had unknown vaccination status.Substandard vaccination compliance is likely to blame for the 2015 measles outbreak, placing the spotlight on our nation’s growing anti-vaccination movement and the prevalence of vaccination-hesitant parents possibly influenced by lively media polemics over the MMR vaccine and its alleged link to autism. If this can occur in a country with a large share of global economic hegemony, capable of providing essential human services, what can one say about the possibilities that may emerge in less developed areas of the world that are rapidly joining the global crowd through technological advances?
What Would a Crowdsourced HUMINT Future Look Like?
We take a bird’s eye view of such areas inhabited by human beings who love and take pride in their cell phone, have started embracing the idea of chatting on Facebook or other social media instead of their traditional community gatherings, and are exposed to the idea of crowdsourcing. Crowdsourcing will facilitate individuals’ access to emergency assistance when needed, maybe even reduce risks to their crops and increase yields, expand trade of products and provide real-time pricing updates, and more benefits on one condition: that people are not only consumers of data, but also contributors to the greater good for this to work. And because to these populations social media serves as a means for showcasing their newly acquired well-being and improved standards of living, we rely on their intrinsic motivationto gain more reputation while making a civic contribution.
On the other side of the screen, these individuals serve as “debriefings” to HUMINT collectors. Like in Little Prince’s human world, we feel like we know them by solely using advanced statistical models to assess their input after which we can throw them into large HUMINT databases. Increasing demand for high speed HUMINT reporting has made data collection and management highly automated, fueling the demand for more advanced biometrics. If we continue along this route, the days of human espionage may be forgotten, computers becoming charged with information vetting and analysis as well as with compiling automated reports of possible imminent threats.
Soldiers’ mind controlling nano-built UAVs will be deployed in areas selected by machines equipped with “human-level intelligence” software after having sifted through a pile of crowd volunteered data. Once Facebook’s Internet.org initiative has managed to provide internet capabilities in remote areas of the world, DARPA’s Warrior Web and TALOS suited soldiers will be able to reach targeted villages without risk of physical harm. The future of HUMINT in this case looks like a combination of computer-based selection of crowdsourced data and area deployment of supermen and possibly superwomen whose enhanced look will only remotely remind us of human beings. Tomorrow’s HUMINT collectors will be left at their desks, alone and waiting to receive “explorers” from other “planets” through crowdsourced information.
Conclusion: Harnessing the Power of Crowdsourcing for Successful HUMINT
Nicholas Mumm provides a tentative new model for HUMINT collection in which a commander can expect the distribution of the population’s participation in a social network built in an area of conflict to follow the power law distribution. If we follow the Pareto Principle (80% of effects come from 20% of the causes) and then profile real, human-generated codes through the power law distribution expecting 80% of the valuable data to be shared by only 20% of the crowd, then it is worthwhile to spend more time developing a close, intimate5 relationship with these 20%.
We go back to the role of HUMINT: “to gather foreign information from people and multiplemediasources to identify adversary elements, intentions, composition, strength, dispositions, tactics, equipment, personnel, and capabilities.” Michael Althoff recently expanded the definition to include both clandestine and overt collection of intelligence by human sources (pp.48). The secret to future successful HUMINT also lies in its own (re)definition of human sources. Clandestine or overt, human sources should not be “used” as “tools,” but empowered as partners. When crowdsourcing becomes “the tool” the crowd utilizes to prevent conflicts, seek or provide help, support, training, etc., then the population is no longer a casualty resulting from political decisions regarding warfare, but an active community contributing to its own political future through the use of technological advances. In this scenario, the crowd is no longer a “fixed institution available on demand,” but a lively enterprise that increases its investment based on bilateral trust and perceived security.
While this approach might seem puerile if considering the incredible technological advances in the areas of nanotechnology, bio-mimetics, and biometrics that will definitely accelerate the race in military superiority, HUMINT still matters in keeping a human approach to warfare to ensure world peace. As increased cell phone usage and expanded internet access have been acknowledged to hold huge potential for counterinsurgent forces to reach to populations of interest by building large social networks, crowdsourcing should have an important place in HUMINT. But mobilizing the crowd through Distributed Human Intelligence Tasking needs to follow the same 80-20% rule, a good relationship with the 20% becoming vital to HUMINT collection efforts.
To ensure a high level of trust and security, HUMINT collectors should follow the “geographer’s” approach: “And if the recollections of any one among them seem interesting to him, the geographer orders an inquiry into that explorer’s moral character.” This would allow the crowd to be beneficial to the extent of its own capability and knowledge, while HUMINT collectors would make important decisions also based on known, field collected (in any form) and documented “morality” of the contributors. The crowd, if properly harnessed and carefully analyzed, has the potential to provide HUMINT with access to accurate and reliable data generated by on-the-ground communities of interest as long as the vetting and analysis processes still involve human input.
Click here to download an annotated PDF of this article.
Camelia M. Kantor, Ph.D., is an assistant professor of geography at Claflin University.
Photo by U.S. Army Capt. Kevin Sandell. Soldiers with the 504th Military Intelligence Brigade conduct human intelligence collection training at an urban training area at Fort Hood, Texas, Sept. 1, 2015.