The Office of the Director of National Intelligence (ODNI), which oversees and directs the work of the 17 agencies and organizations responsible for foreign, military, and domestic intelligence for the United States, has an interest in research from the social and behavioral sciences that may be beneficial to the Intelligence Community (IC). To develop a systematic understanding of these potential benefits, ODNI requested that the National Academies of Sciences, Engineering, and Medicine conduct a decadal survey of the social and behavioral sciences to identify research opportunities that show promise for supporting national security efforts in the next 10 years.
A decadal survey is a method for engaging members of a research community to identify lines of research with the greatest potential utility in the pursuit of a particular goal. The National Academies pioneered this type of survey with a study of ground-based astronomy in 1964.1 Since then, committees appointed by the National Academies have conducted more than 15 decadal surveys. The Decadal Survey of Social and Behavioral Sciences for Applications to National Security represents the first opportunity to apply this approach to the social and behavioral sciences. Its purpose is to develop an understanding of the lines of research in these
fields that offer the greatest potential to enhance the capabilities of the IC. To carry out this work, the National Academies appointed the Committee on the Decadal Survey of Social and Behavioral Sciences for Applications to National Security (Decadal Survey Committee); the committee’s charge appears in Appendix A.
The Decadal Survey Committee has pursued many avenues in collecting information about the needs of the IC and relevant cutting-edge research in the social and behavioral sciences. As part of its information-gathering process, the committee held a series of six workshops—the first three on October 11, 2017, and the second three on January 24, 2018.2 These workshops, for which planning began early in the committee process, were designed to explore areas about which the committee wished to learn more and to allow the committee to engage with a broad range of experts. The topics selected for the workshops do not necessarily indicate the ultimate direction of the committee’s deliberations. The six topics addressed by the workshops were
- changing sociocultural dynamics and implications for national security;
- emerging trends and methods in international security;
- leveraging advances in social network thinking for national security;
- learning from the science of cognition and perception for decision making;
- workforce development and intelligence analysis; and
- understanding narratives for national security purposes.
Separate steering committees, whose membership included both members of the Decadal Survey Committee and additional experts in the topics to be addressed, were appointed to plan these workshops. Each of these committees was guided by its own charge. All were asked to bring their expertise to bear in identifying specific areas of promising research and experts with deep knowledge who could offer a range of insights.
This Proceedings of a Workshop, prepared by the workshop rapporteur, summarizes the presentations and discussions at the third workshop, which addressed advances in social network thinking.3 This workshop was planned by the Steering Committee on Understanding Networks for National Security Purposes, whose charge is presented in Box 1-1. The workshop’s purpose was to explore the current state of research on social
3 For the archived Webcast of the workshop and available presentations, see http://sites.nationalacademies.org/DBASSE/BBCSS/DBASSE_181267 [November 2017].
network thinking that has relevance for national security. It should be noted that the steering committee’s role was limited to planning and convening the workshop, and that the views contained in this proceedings are those of individual workshop participants and do not necessarily represent the views of all workshop participants, the steering committee, or the National Academies. The agenda for the workshop appears in Appendix B; a list of individuals who attended the three workshops held on October 11, 2017, is presented in Appendix C; and biographical sketches of the steering committee members and speakers are provided in Appendix D.
In an opening session for the three October 11, 2017, workshops, the chair of the Decadal Survey Committee, Paul Sackett, University of Minnesota, and sponsor representative David Honey, ODNI, provided background information on the objectives for the six workshops.
Sackett observed that the Decadal Survey Committee will rely heavily on input from experts in the communities of national security and behavioral and social science research. Given the breadth of the committee’s charge, he explained, it must cast a wide net, extending well beyond the specific expertise of its members. He described the six workshops as an important part of the effort to gather ideas. The workshops would support the committee by helping to identify promising research areas and allow-
ing the committee members to engage in discussion with experts in a wide range of areas salient to its work.4
Honey expressed appreciation to all those contributing to the committee’s work through the workshops and other activities, noting that the participation of the full range of experts in the intelligence and behavioral and social science communities would be needed to make the decadal study successful. Making predictions about future directions for research is difficult, he acknowledged, but in his view it is necessary. He noted that the final report of the Decadal Survey Committee will be “a very powerful tool” for government officials who must make decisions regarding funding and other priorities. The decadal model, he explained, “offered the best opportunity” to identify research directions and priorities that reflect a wide range of insights and perspectives. “Decision makers are really asking much deeper and more probing questions today than we’ve seen before,” he said. “They really want to know why surprising movements such as the Arab Spring [uprisings that began in 2010] occur. The national security community is eager for new ways to understand such events and how to respond to them, and also for better ways to assess their interventions after the fact.” Honey thanked the participants for contributing, emphasizing that their ideas would be “crucial for getting us where we need to go.”
The workshop steering committee recognized that advancing technology is changing the environment in which humans reside, and that these changes are in turn influencing the way social networks form and evolve, as well as how they might be studied. Scholars in many disciplines have developed theories and research strategies to pursue understanding of these changes, and also to utilize an unprecedented availability of data. Researchers are applying their evolving approaches to address a range of social phenomena and problems. To explore this range of work, the committee invited presenters to describe research that applies social network thinking to several areas of inquiry. Presenters were instructed to consider recent advances in their areas of work and the gains that could be made with future investment.
Committee chair Kathleen Carley, Carnegie Mellon University, explained that a primary goal for the workshop was to explore ways to “reason about and understand relations among things, people, organiza-
tions, and ideas” and how that understanding can support intelligence decisions. She provided an overview of social network thinking and some of the challenges related to methods and data used in studying networks. She noted that the science of social networks has expanded: the tools and data sources available today are quite different from those of the past. The field has relied on surveys of small groups of people that identify the various relationships within an organization or community, and those methods are still used, she acknowledged, but attention has shifted to ever larger datasets and networks.
Carley added that new technologies, such as those developed for automated communications, machine learning, psychological profiling, speech analysis, and video modification, are changing the ways in which social networks are formed and investigated. However, she asserted, these advances are not without challenges. One such challenge is that these technological advances are also increasing opportunities for “enhanced disinformation” about what is really going on within a given network and among whom. For example, Carley noted, advances in voice and image software are opening up the possibility of generating fabricated images and videos of conversations between people that appear to be real, even though they never occurred. She added that machine-driven bots or cyborgs (humans who use bot technology) can create the illusion of human interaction in online networks, as well as introduce propaganda into online discussions.5 She noted that producing credible analyses of social networks when the underlying interaction data may be suspect is a challenge.
Carley introduced three terms used within the community of researchers on social network thinking: “echo chamber,” “superspreader,” and “superfriend.” An “echo chamber” refers to a network of actors who are densely connected in at least two ways: they are in frequent communication with each other, and they use shared terms or address a shared set of topics. Within echo chambers, information, ideas, and beliefs diffuse rapidly and are amplified and reinforced within the network. A “superspreader” is someone who can send information widely to numerous individuals, directly or indirectly. When an idea is mentioned by a superspreader, it is likely to be “heard” by many others. Finally, a “superfriend” is someone who is involved in numerous reciprocal relationships and is critical in engaging groups through ongoing dialogue. Carley noted that these three
5 Carley gave the example of the FiribiNome bot to illustrate why new network thinking tools are needed and being developed. One of the echo chamber bots associated with ISIS could not be discovered using traditional social network metrics. Within the Twitter platform, sophisticated use of @mentions through the bot made it possible to grow networked communities, promote accounts, and gain influence on these communities. Once the bot had manipulated Twitter recommendations and grown a community of followers, it promoted a site for the collection of money for children of Syria that was actually a money-laundering site for ISIS.
concepts have been in use for some time and that researchers have validated ways of measuring them. She pointed out that these concepts are important to the types of social networks observed today, such as those in social media. However, she observed, the traditional measures may not be suited to analysis of new data on networks given the volume of the data and their time-variant nature, and more complex dynamic metrics are needed.
Accordingly, Carley reported, the science of social networks today is making use of data from remote sensors and such techniques as machine learning, big data analytics, and neural imaging to investigate people’s behaviors within social networks. She explained that the social media landscape has created new types of networks to which vast amounts of data are tied, and researchers are eager to use this wealth of data to advance social network thinking. However, the challenges she had alluded to earlier complicate what questions can be investigated with these data and what can be known about these social networks. In addition to the challenges previously cited, she observed that sampling biases may be introduced when companies or data providers limit access to some data and when legal authorities limit the collection and storage of certain data.
Carley explained that the workshop would consider two emerging notions of the science of social networks that have come to be referred to as “networks-plus” and “complexity.” The term “networks-plus” refers to research that goes beyond the simplified view of “who-to-whom” connections in a network to consider the contexts associated with or external influences on that network, as well as the contributions of physiological, psychological, and cognitive factors that influence the individual network members. The term “complexity,” she said, refers to two additional approaches: (1) the study of new forms of large, global online networks; and (2) the effort to use network thinking for forecasting and predicting future actions, rather than just for purposes of description and comparison.
This proceedings follows the structure of the workshop. Chapter 2 summarizes four workshop presentations on the external influences on networks (an aspect of networks-plus research). These presentations covered assembling analytic teams of supersynthesizers,6 team organization, urban network research, and the use of crowdsourced data. Chapter 3 turns to research on the individual and internal factors of networks (another aspect of networks-plus research). It summarizes two presentations on the relation
6 As described in Chapter 2, supersynthesizers are people with cognitive and analytic skills beyond those of “superforecasters,” a term that refers to people identified as the most accurate in determining future geopolitical and economic outcomes.
of brain functioning and social networks and two other presentations on emotional artificial intelligence and the study of social identities. The three presentations summarized in Chapter 4 addressed complexity in terms of “multilevel, high-dimensional, evolving, and emerging networks,” examining the exploration of dark networks, robust summary statistics for networks, and the future of complex networks. Finally, Chapter 5 summarizes the discussions between presenters and other workshop participants on the topics raised during the workshop, focusing on collaboration networks in the IC, areas for further research on networks and network thinking, data needs, and ethical considerations in conducting research on social networks.
This page intentionally left blank.