Artificial intelligence for wargaming and modeling

Abstract

In this paper, we discuss how artificial intelligence (AI) could be used in political-military modeling, simulation, and wargaming of conflicts with nations having weapons of mass destruction and other high-end capabilities involving space, cyberspace, and long-range precision weapons. AI should help participants in wargames, and agents in simulations, to understand possible perspectives, perceptions, and calculations of adversaries who are operating with uncertainties and misimpressions. The content of AI should recognize the risks of escalation leading to catastrophe with no winner but also the possibility of outcomes with meaningful winners and losers. We discuss implications for the design and development of families of models, simulations, and wargames using several types of AI functionality. We also discuss decision aids for wargaming, with and without AI, informed by theory and exploratory work using simulation, history, and earlier wargaming.