Multi-Agent Strategic Games with LLMs

2026-05-05

AI Summary

This study investigates whether large language models (LLMs) can reproduce the strategic mechanisms of conflict and cooperation central to international relations theory. By deploying LLMs as agents in repeated security dilemma games, the research systematically examines how multipolarity, finite time horizons, and communication availability shape strategic behavior. Through analysis of both private reasoning traces and public action logs, the authors uncover the underlying decision-making logic. The work introduces a scalable, transparent, and reproducible multi-agent experimental framework and demonstrates that multipolarity intensifies conflict, limited temporal horizons undermine cooperation, and communication significantly enhances cooperative outcomes through signaling and reciprocity. These behavioral patterns are successfully linked to canonical strategic logics, including preemptive action, cooperation under uncertainty, and trust formation.

Abstract

This paper asks whether large language models (LLMs) can be used to study the strategic foundations of conflict and cooperation. I introduce LLMs as experimental subjects in a repeated security dilemma and evaluate whether they reproduce canonical mechanisms from international relations theory. The baseline game is extended along three theoretically central dimensions: multipolarity, finite time horizons, and the availability of communication. Across multiple models, the results exhibit systematic and consistent patterns: multipolarity increases the likelihood of conflict, finite horizons induce universal unraveling consistent with backward-induction logic, and communication reduces conflict by enabling signaling and reciprocity. Beyond observed behavior, the design provides access to agents' private reasoning and public messages, allowing choices to be linked to underlying strategic logics such as preemption, cooperation under uncertainty, and trust-building. The contribution is primarily methodological. LLM-based experiments offer a scalable, transparent, and replicable approach to probing theoretical mechanisms.

Problem

• Multi-Agent Strategic Games
• Large Language Models
• Security Dilemma
• International Relations Theory
• Strategic Behavior

Innovation

• LLM-based experiments
• strategic reasoning
• security dilemma
• multi-agent simulation
• international relations theory