Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Open-Ended Wargames with Large Language Models

Daniel P. Hogan, Andrea Brennen

TL;DR

The paper tackles automating open-ended qualitative wargames, which have been difficult to automate with prior quantitative-focused work. It introduces Snow Globe, an open-source Python-based multi-agent framework where a control moderator, AI/human players, and teams interact through a history object and LLM-powered prompts to produce coherent narratives and adjudications. The authors detail the architecture, a zero-shot, locally runnable LLM pipeline, and an asynchronous human-in-the-loop design, plus two demonstrations: an AI incident-response tabletop exercise and a geopolitical crisis simulation with persona-based dynamics. The work argues that unconstrained, text-based moves become tractable with LLMs, enabling rapid, diverse scenario exploration and human-AI collaboration across domains.

Abstract

Wargames are a powerful tool for understanding and rehearsing real-world decision making. Automated play of wargames using artificial intelligence (AI) enables possibilities beyond those of human-conducted games, such as playing the game many times over to see a range of possible outcomes. There are two categories of wargames: quantitative games, with discrete types of moves, and qualitative games, which revolve around open-ended responses. Historically, automation efforts have focused on quantitative games, but large language models (LLMs) make it possible to automate qualitative wargames. We introduce "Snow Globe," an LLM-powered multi-agent system for playing qualitative wargames. With Snow Globe, every stage of a text-based qualitative wargame from scenario preparation to post-game analysis can be optionally carried out by AI, humans, or a combination thereof. We describe its software architecture conceptually and release an open-source implementation alongside this publication. As case studies, we simulate a tabletop exercise about an AI incident response and a political wargame about a geopolitical crisis. We discuss potential applications of the approach and how it fits into the broader wargaming ecosystem.

Open-Ended Wargames with Large Language Models

TL;DR

The paper tackles automating open-ended qualitative wargames, which have been difficult to automate with prior quantitative-focused work. It introduces Snow Globe, an open-source Python-based multi-agent framework where a control moderator, AI/human players, and teams interact through a history object and LLM-powered prompts to produce coherent narratives and adjudications. The authors detail the architecture, a zero-shot, locally runnable LLM pipeline, and an asynchronous human-in-the-loop design, plus two demonstrations: an AI incident-response tabletop exercise and a geopolitical crisis simulation with persona-based dynamics. The work argues that unconstrained, text-based moves become tractable with LLMs, enabling rapid, diverse scenario exploration and human-AI collaboration across domains.

Abstract

Wargames are a powerful tool for understanding and rehearsing real-world decision making. Automated play of wargames using artificial intelligence (AI) enables possibilities beyond those of human-conducted games, such as playing the game many times over to see a range of possible outcomes. There are two categories of wargames: quantitative games, with discrete types of moves, and qualitative games, which revolve around open-ended responses. Historically, automation efforts have focused on quantitative games, but large language models (LLMs) make it possible to automate qualitative wargames. We introduce "Snow Globe," an LLM-powered multi-agent system for playing qualitative wargames. With Snow Globe, every stage of a text-based qualitative wargame from scenario preparation to post-game analysis can be optionally carried out by AI, humans, or a combination thereof. We describe its software architecture conceptually and release an open-source implementation alongside this publication. As case studies, we simulate a tabletop exercise about an AI incident response and a political wargame about a geopolitical crisis. We discuss potential applications of the approach and how it fits into the broader wargaming ecosystem.
Paper Structure (11 sections, 2 figures, 4 tables)

This paper contains 11 sections, 2 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Quantitative and Qualitative Wargames
  3. Automated Wargames
  4. Software Architecture
  5. Conceptual Approach
  6. Software Implementation
  7. Human-Machine Interaction
  8. Example Simulations
  9. Tabletop Exercise
  10. Geopolitical Crisis Simulation
  11. Conclusion

Figures (2)

  • Figure 1: Taxonomy of human and automated wargames.
  • Figure 2: Software architecture examples for the Snow Globe multi-agent system. Vertical arrows represent agents and horizontal arrows represent information flow for (a) a simple tabletop exercise and (b) a team of players formulating a collective response.