Multi-Agent Strategic Games with LLMs
Maxim Chupilkin
Why It Matters
What makes this one worth your time
Understanding how LLMs can simulate strategic interactions offers a novel approach to studying complex social dynamics and international relations theories.
LLMs are used to simulate strategic games, revealing insights into conflict and cooperation dynamics.
Summary
The paper explores the use of large language models (LLMs) as experimental subjects in strategic games to study conflict and cooperation, extending the baseline game with multipolarity, finite time horizons, and communication. It finds that these extensions affect conflict likelihood and strategic behaviors, providing insights into strategic logics like preemption and trust-building.
Key contributions
- Introduces LLMs as experimental subjects in strategic games.
- Extends strategic games with multipolarity, finite horizons, and communication.
- Links LLM behavior to strategic logics like preemption and trust-building.
Notable insights
- LLMs can simulate strategic reasoning in multipolar scenarios.
- Communication among LLMs reduces conflict through signaling and reciprocity.
Possible limitations
- Not stated in the abstract
Abstract
arXiv:2605.03604v1 Announce Type: cross 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.