2
Opportunities for the Intelligence Community
The committee explored both the needs of the IC and the areas of SBS research to identify key opportunities to strengthen intelligence analysis and national security (see Box 4 for a description of their process). They identified significant research opportunities in four key areas:
- sensemaking: emerging ways to answer intelligence questions;
- enhancing security in cyberspace;
- supporting the design of a human–machine ecosystem; and
- strengthening the analytic workforce for future challenges
For each of these areas the committee identified specific ways the IC could benefit from research developments that can be reasonably expected in the coming decade, if priority is placed on supporting this work. The committee did not claim that these are the only areas of opportunity, but was confident that they are “ripe” in the sense that they offer innovations in theory and/or application likely to bear fruit in concrete ways in the coming decade, and are responsive to significant goals and needs related to the analyst’s work. The full report provides detailed discussion of possible applications and research directions to support progress in each area.
SENSEMAKING: EMERGING WAYS TO ANSWER INTELLIGENCE QUESTIONS
Understanding of human social processes—insights about the functioning of individuals, groups, and societies—is essential to the analysts’
ability to answer enduring intelligence questions. Analysts’ primary function is sensemaking, drawing meaningful conclusions from the vast stream of information to which they have access about core phenomena related to national security, including the nature of power and influence; threats, opportunities, and social and organizational dynamics; complexity; and deception and gaps in information. Figure 1 suggests the complexity of the challenge of making sense of diverse kinds of information from varied sources that may be relevant to an analyst’s primary area of responsibility.
Four areas of SBS research have the potential to be particularly fruitful in supporting analysts’ sensemaking capacities.
The Study of Narrative
Understanding the content of communications and how and why they are conveyed—from the meaning of cultural traditions, to political themes in press coverage, to trends in social media communications—is fundamental for the intelligence analyst. Scholars in in the humanities and in such social sciences as anthropology and psychology have long studied the structure and content of narratives using many tools and methods. More recently, exponential growth in data created by social media such
SOURCE: Generated by the Committee on a Decadal Survey of Social and Behavioral Sciences for Applications to National Security.
as Facebook and Twitter, and by the fact that vast amounts of content are now stored digitally, has made the study of narratives at a large scale much more practical. These capabilities offer new frontiers for applying the study of narratives to intelligence analysis.
Among the possibilities are decoding influential narratives such as those of Islamist extremism, interpreting new kinds of data such as digital video footage, and tracking the flow and influence of ideas or emotional states. Techniques for analyzing multiple aspects of communication and its context can yield insights into the comparative power and influence of political narratives and means of countering them. Machine learning techniques have opened up possibilities for developing effective indicators of growing extremism or potential for violence in narrative streams.
The Study of Social Networks
Social network analysis is a structural approach to understanding the world based on the interdependencies among actors and their influences on behavior. This type of analysis has played an important role in fields such as anthropology, communication, sociology, and political science. It entails representing a network in terms of nodes and relations that form an interdependent, holistic system, and identifying key actors, their group identifications, and other network features. Cutting-edge methods for social network analysis rest on technological advances, particularly improved capacity for application at very large scales. Many of these advances resulted from collaboration between researchers and the IC—but the utility of this research for intelligence analysis rests on interdisciplinary work with other SBS disciplines.
Social network analysis can yield understanding of such phenomena as the distribution and exchange of resources, the development of trust within a group or society, ideological contagion, diffusion of beliefs and attitude formation, the establishment of normative constraints, the development of group and individual social capital, group and organizational effectiveness, the evolution of organizational leadership, group and organizational resilience and robustness, and political stability, among many others.
The Study of Complex Systems
Subject matter experts in the IC face constant pressure, often within tightly compressed timeframes, to understand and develop forecasts about scenarios that are both complex and constantly changing. Many of the issues they follow have the features of complex systems. A good example is the role of China in the world, which is a key issue for the IC. Numerous factors—including economic performance, degree of social cohesion,
rural-urban migration, leadership dynamics in the Communist party, and environmental degradation, to name but a few—interact and shape developments in that country. Yet analysts are expected to communicate findings about China to policy and decision makers in clear terms.
Work in fields including both mathematics and philosophy has contributed to the development of an interdisciplinary approach to studying complexity (sometimes termed complexity science, strategy, or theory). Based in systems theory as well as developments in the natural sciences, this approach is used to study phenomena that are unpredictable and nonlinear, providing ways to identify and mitigate unintended consequences, as well as methods useful for considering a wide range of alternatives and thus supporting strategic analysis—all of which have clear utility of intelligence analysis. Scholars of complexity theory use computational and mathematical methods to assess such phenomena, and rely heavily on modeling and simulation.
The Affective Sciences
The affective sciences are fields of study that address emotions, feelings, affect, moods, sentiments, and affectively based personality traits and psychopathologies. The study of these kinds of phenomena, including the verbal and nonverbal signals of affective states, can provide insights into the mindsets, personalities, motivations, and intentions of the actors whom intelligence analysts seek to understand; help explain people’s actions, judgments, and decisions; and support more nuanced and sophisticated understanding of communication. Research has provided strong support for the validity and reliability of interpretations of nonverbal expressions of specific emotional states and signals of other cognitive and emotional states. Such findings can support the analysis of phenomena including the content and power of narratives, the processes of judgments and decision-making, and the spread of attitudes and beliefs associated with terrorism and security threats.
Basic research in these four areas provides a theoretical and empirical foundation for the development of sophisticated methods that analysts can use in tackling core sensemaking challenges. For example, accurate computer modeling of complex, sophisticated sociopolitical systems rests on foundational understanding of the nature of status and power, socioemotional processes, and linguistic structures. The potential for advances in the development and application of such tools and approaches, in turn, rests on the marriage of technological advances and insights from SBS fields. Examples include the use of digital trace data to assess the importance of network nodes that pose potential security threats or to track the trajectory of political ideas, the use of understanding of how nonverbal cues can
Sophisticated methods such as computational analysis of large datasets would make little sense without theoretical frameworks to guide the development of algorithms, such as those for classification of narrative structures or analysis of the functioning of social networks. Similarly, insights from SBS research are important guides for the development of indicators that could be used to track, for example, significant emotional states or changes in leaders or other powerful actors, the developing strength of a minority group’s message, or the cohesiveness of networks in which toxic narratives are spreading.
There is considerable variation in how directly SBS research has been applied to questions of interest to the IC, in how close it is to providing the basis for practical application for analysts, and in the aspects of intelligence analysis it could potentially support. Tools for social network analysis, for example, are close to being operational, and simulation groups within the IC are using models to address complex situations. Research on other tools, such as those based on understanding of unconscious behavior and nonverbal cues, is still emerging.
Further progress in the development of applications that can serve the IC’s needs will depend on interdisciplinary collaboration that takes advantage of developments in multiple fields. For example, the integration of recent advances in analysis of narratives, networks, and affect would provide a framework for supporting dramatic advances in the assessment of narratives and counternarratives, early detection of polarization, assessment of group vulnerability to disinformation, and detection and mitigation of diverse information maneuvers. Similarly, applying multidisciplinary, multimethods research to IC issues from a complexity perspective—using modeling and simulation, representation, and understanding of human factors—would yield significantly stronger methods of forecasting surprising events or developments.
Advances in the use of large-scale data are likely to be at the heart of significant developments for the IC in the coming decade, but new technologies will be only as strong as the understanding of the human phenomena they are used to model or explain. The committee expects that there will be progress in the development and validation of computational models, the reuse of simulation modes, and the integration of social networks with computational models. These advances have the potential to make near-real-time assessment of competing actors, messages, or groups and the interventions that influence them possible.
INTEGRATING SOCIAL AND BEHAVIORAL SCIENCES (SBS) RESEARCH TO ENHANCE SECURITY IN CYBERSPACE
Cyber-related developments have both dramatically altered the nature of security threats and expanded the landscape of potential tools for countering those threats. Experts from multiple disciplines, including electrical engineering, software engineering, computer science, and computer engineering, have a laser focus on cybersecurity, but those efforts have primarily addressed technical or data challenges: protecting the integrity of networks, programs, and data. These techniques have undisputed value but they shed relatively little light on the human behaviors and motivations that shape cyber-based challenges.
The emerging field of social cybersecurity science has developed to fill the need to integrate understanding of constantly evolving technology with insights about fundamentally human phenomena. Researchers in this field build on foundational work in SBS fields to characterize cyber-mediated changes in individual, group, societal, and political behaviors and outcomes, and also to support the building of the cyber infrastructure needed to guard against cyber-mediated threats.
Designing ways to protect against cyber-based threats requires the ability to collect data on and analyze and visualize high-dimensional dynamic networks with both social network and knowledge network components. Twitter networks, for example, generate both social data on who replies, retweets, or mentions or which individuals are quoted, and knowledge data on hashtags or topics that co-occur. However, available machine learning techniques and standard computer science methods are of limited utility for answering nuanced questions about developing situations. Nor are traditional social science methods sufficient to address complex issues in today’s information environment.
A promising next frontier is the combining of computer science techniques with deep theoretical understanding—from social and cognitive science research—of such phenomena as how the media and entertainment technology used to collect these data operate and the nature of the sociocultural phenomena being studied. Methods used in network science, coupled with language technologies, geospatial crowdsourced information, or machine learning and applied to large-scale data form the methodological cornerstone on which further advances social cybersecurity will be realized.
Empirical assessment of influence and manipulation in social cyberspace is yielding methods capable of processing large volumes of data, often from multiple media, and carrying out high-dimensional network analysis. Such methods have been successfully used to address a number of issues, such as the likelihood of retweeting, information diffusion, disaster planning, extremist recruiting, and political polarization. Furthermore, geo-
spatial assessments have shown great diversity in the ways in which social media are used by region, time, and political context. This work provides a starting point for the development of tools that could be used by the IC for efficiently identifying propaganda, false information, and other social cyberthreats.
INTEGRATING THE SOCIAL AND BEHAVIORAL SCIENCES (SBS) INTO THE DESIGN OF A HUMAN–MACHINE ECOSYSTEM
Technologies that become operational in the coming decade and beyond will be capable of augmenting the capacities of the human in vital ways, and these developments will necessarily change the ways human analysts use and interact with the technological resources available to them. Figure 2 illustrates the sort of human-machine ecosystem that could become a central component of intelligence analysis.
Insights from SBS fields are essential to the design and development of tools and technologies that
- take advantage of the strengths of both humans and machines;
- allow humans to collaborate productively with machine partners;
- support more rapid assessment and forecasting of human activity; and
- avoid serious unintended practical and ethical consequences.
SBS research offers insights on human capacities and limitations, how humans can interact effectively with machines, how humans and machines can collaborate as teams, and how machines can mimic and manipulate humans. These insights will be needed in the design of tools that use Artificial Intelligence (AI) and machine learning in conjunction with social network analysis, which is likely to be an increasingly important component of analysis. This work could also support the development of an ecosystem for intelligence analysis composed of human analysts and semiautonomous AI agents, operating on and through diverse social media and supported by other technologies. Such a team could, proactively and securely, reach across controlled-access networks and develop enhanced intelligence analyses by identifying patterns and associations in data more rapidly than humans alone could, doing so in real time and uncovering connections that previously would not have been detectable.
Whatever directions the IC takes in developing and procuring technologies to support intelligence analysis in the coming decade, it will surely rely on researchers and other experts, both those working within the IC and outside contractors; commercially available software programs; and other resources. The extent to which both basic and emerging SBS research is already being incorporated into the planning, design, and use of the tools and methods used and purchased by the IC is not publicly known. Emphasis on this aspect of design is critical, however, because the technology used for analysis is only as strong as the understanding of the human behavior it is being used to model or explain; insights from SBS fields will provide essential support for the procurement of valid and effective products from the private sector to support the analyst’s work.
NOTE: This figure is a simple illustration of a human–machine ecosystem with analysts working in collaboration with and connected through AI systems. It shows three teams of analysts who have different sets of responsibilities working individually while remaining connected to the ecosystem. A variety of sensors (shown surrounding the ecosystem) provide information to the human and AI agents for a number of different purposes, from monitoring and analyzing data pertinent to intelligence analysis to collecting and processing data from interactions between analysts and AI to improve performance.
SOURCE: Generated by the Committee on a Decadal Survey of Social and Behavioral Sciences for Applications to National Security.
For some, talk of machines and collaboration with AI agents can be somewhat chilling. Reasonable concerns include the prospect that, in restricted environments, there will be more opportunities for inadvertent disclosures of confidential information; that biases inherent in algorithms will negatively affect decision making; that machine-generated output may increase false positives and subsequent false alarms; and that too much trust may be placed in machines to find the emergent patterns and signals, perhaps usurping what should be functions of human analysts or occupying them with new oversight and management tasks that compete with their analytic work.
STRENGTHENING THE ANALYTIC WORKFORCE FOR FUTURE CHALLENGES
The IC has always needed a workforce that is responsive, flexible, effective, and well equipped to learn and adapt to change, but technological developments are likely to bring fundamental changes in the way intelligence analysis is conducted. In a decade or less, for example, analysts may have the capacity to obtain sophisticated analysis of a months-long narrative stream on social media sites, compare it with activities from that period identified through geospatial imaging, and develop a graphical representation of the intersections between the two—as part of a day’s work.
To take advantage of these opportunities, the analytic workforce will need new skills: developments in such areas as network science, complex systems models, statistics, and data analytics of all kinds will likely add new methods and tools to the analyst’s toolbox. In areas in which intelligence analysts are expert—qualitative analysis of text and narrative, for example—new developments such as improved quantitative methods for text analysis, including methods for analyzing social media, offer possibilities that may not yet have been integrated into common practice within the IC.
The analytic workforce already reflects diverse and valuable technical and academic skills and experience, and analysts typically join the workforce with specific disciplinary subject matter knowledge. Analysts of the future will need to build on the skills they have always had, including technical skills, domain-specific knowledge, social intelligence, strong communication skills, and the capacity for continued learning—but they will also need to function in new ways.
As in any large organization, the agencies of the IC pay attention to means of identifying, recruiting, and selecting individuals likely to excel as intelligence analysts; providing training and using other means to develop their skills and abilities; obtaining the best possible performance from the workforce; and retaining effective employees. Researchers in the fields of industrial-organizational psychology and human resource development have produced a robust body of work on ways to pursue most of these objectives.
Translational research is needed, however, to identify specific ways the IC can take advantage of these opportunities. Moreover, the nature of analytic work is a moving target. In the next 10 years, new work challenges, new analytic technologies, and new work practices can be expected to emerge, and new collaborations will become necessary. Selection practices, training regimes, and teamwork requirements will need to be adapted to the new work requirements that will result.
This page intentionally left blank.
