Deus Ex Machina? Questions from Wargaming
Abstract: Since the nineteenth century, wargaming has been a component of planning and training in military contexts – and its techniques and concepts have spread to many other arenas. Traditional wargaming has had a cyclical relationship with technical innovation in the national security community, first compensating for gaps, then being largely supplanted, and in the last forty years coming to focus heavily on the human element of an analytic or decision-making cycle. With the meteoric rise in interest around LLM-based AI, the defense establishment is focusing attention on how these technological innovations may support wargaming, break it, or enhance it. Modeling and simulation have always been an integral part of the wargaming enterprise, but their effectiveness and proper role remains a key matter of debate, this is especially true around questions of how much current tech could replace humans in certain scenarios, and how much non-technical end users understand the strengths and weaknesses of the product that they are looking at. This presentation attempts to explore the critical questions in this space and offer some insight into current practices for those unfamiliar with wargaming methodology and practice. It will be highly focused on conversation and eliciting insight.
Bio: Devin Hayes Ellis is a senior researcher at the Applied Research Lab for Intelligence and Security at the University of Maryland. He is a policy analyst by training who began his career in U.S.-China defense analysis and fell backwards into running a wargaming organization. He has designed and executed wargames supporting every Combatant Command, four Services, the Joint Staff, OSD, numerous Component Commands, the State Department, USIP, and various parts of the Intelligence and Law Enforcement communities. In his current post as head of the Asymmetric Threats Analysis Center, he is leading a broad assessment of the relationship between AI and M&S and wargaming for the Basic Research Office in OSD/RDT&E.
