The Role of Information in Structured Conflict

Abstract

Conflict is cognitive. When individuals and groups find themselves in conflict, they use a variety of strategies to decide how, when, and who to attack, how to respond to attacks from others, and how to allocate scarce resources to competing tactical demands. At the heart of this process is the gathering and processing of social information: individuals make critical decisions on the basis of perceptions of others. The course of conflict is defined by how individuals synthesize social information and leverage that information for their own gain. Social environments are messy and individuals must infer general patterns to make decisions from the behavior they experience and observe, under potentially cognitively-demanding conditions. Conflict is driven by what individuals know, how they know it, and what that information makes them do. These aspects need not be studied in isolation. Our proposal can examine them together and provide a generalizable theory of the role of information in structured conflict. We propose a series of networked, behavioral experiments to probe the complex feedbacks between information and social order in conflict. Groups of ten to twenty individuals will be placed in zero-sum conflict in a simple virtual environment. They will battle for status in the presence of precisely-calibrated information about the consequences of their own behavior and the behavior of others. They will encounter artificial agents that probe, manipulate and outperform human opponents. Throughout, we will track, second-by-second, the strategies players invent and discard. We will measure how their social environments increase in complexity through the interaction between perception, action, and information, at both the individual and group level. The role of information in conflict has three aspects that make its study particularly difficult. (1) Social information drives conflict. Knowledge about relative status, position in a hierarchy, or membership in an alliance, inferred from messy data, changes what individuals do. (2) Conflict drives social complexity. Individuals do not simply react to a fixed environment. When they gather information, and use it to choose strategies, they dynamically create new facts and alter old ones: relative status might change, new hierarchies might form and old ones collapse, alliances might strengthen or break apart. (3) Social information may be highly dispersed. Information that guides conflict is both local and global. Information flows not only between individuals, but also between the individual and the group. Whether one individual has higher status than an opponent, for example, is determined by the relative status of many others. The feedback between information gathered and action taken can make conflict look chaotic and fundamentally unpredictable. The fact that information can exist at multiple levels-from the individual to the group, and at all stages in-between-makes the problem even harder. Our approach overcomes these difficulties via simultaneous tracking and manipulation of information at multiple scales, and through focusing on the feedback between perception and action. Our approach is new, but the questions we intend to answer have been at the core of the science of conflict for over a century, and our work is of more than academic interest. Antagonists in the modem era often have high levels of internal conflict, where ambitious individuals can gain status and prestige through strategic, but selfish, action. Our proposal isolates this phenomenon from the larger problem of the social dilemma in inter-group conflict. It allows us to focus on the interactions of decentralized antagonists as they develop internal hierarchies and alter their strategies to gain power and achieve objectives. If we understand how individuals convert social knowledge to strategic decisions, we can better predict and manipulate interpersonal conflict.

Document Details

Document Type
DoD Grant Award
Publication Date
May 07, 2018
Source ID
W911NF1710502

Entities

People

  • Simon Dedeo

Organizations

  • Army Contracting Command
  • Massachusetts Institute of Technology
  • United States Army

Tags

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Strategic Security Studies
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy