Shall We Play a Game? Language Models for Open-ended Wargames

2025-09-21

AI Summary

This study addresses the high-uncertainty, creativity-intensive domain of open-ended wargaming, where language models (LMs) face challenges in supporting player strategic decision-making and adjudicator-driven dynamic scenario resolution. Method: We propose the first LM application framework tailored to open-ended wargaming, introducing a three-layer creative dimension taxonomy—intention generation, rule negotiation, and narrative evolution. Through a scoping review of 127 interdisciplinary studies (spanning AI, NLP, and game design), we identify six core technical challenges: interpretable modeling, multi-agent coordination, safety alignment, among others, and formulate military-simulation–aligned ethical safety guidelines. Contribution/Results: The work yields a deployable implementation guide and empirically grounded best practices, establishing both theoretical foundations and actionable technical pathways for integrating LMs into strategic wargame simulation and immersive entertainment applications.

Technology Category

• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Computational Creativity
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

Wargames are multi-faceted, multi-player depictions of conflict in which participants' decisions influence future events. Wargames are often used to explore the strategic implications of decision-making. However, it also encompasses entertainment-oriented simulations, ranging from _Chess_ to tabletop role-playing games like _Dungeons & Dragons_ (D&D). On the more open-ended side of the spectrum of wargames, players use natural language to convey their moves, and adjudicators propose outcomes. Language Models (LMs) are increasingly being considered for how they can provide insights into real-world, consequential decisions. We conduct a scoping literature review of a curated selection of 100 recent works on AI in wargames, from which we construct an ontology of wargames in terms of the creativity afforded to either the players or adjudicators. Focusing on the space of wargames with the most open-endedness for players and adjudicators, we distill a set of considerations for when and how to use LMs in different application areas. We also present a set of safety considerations, best practices for deploying LMs in open-ended wargames, and conclude with a set of high-impact open research challenges.

Problem

• Exploring how Language Models can provide insights into real-world strategic decision-making through wargames
• Constructing an ontology of wargames based on creativity afforded to players and adjudicators
• Developing safety considerations and best practices for deploying LMs in open-ended wargaming scenarios

Innovation

• Conducting scoping literature review
• Constructing ontology of wargames
• Distilling LM deployment considerations