When Identity Overrides Incentives: Representational Choices as Governance Decisions in Multi-Agent LLM Systems
This work addresses governance challenges in deploying LLM systems for strategic tasks, revealing that representational choices critically influence behavior, which is incremental but important for AI safety and policy applications.
The study investigated how design choices like role-based personas and payoff visibility affect LLM agents in multi-agent strategic games, finding that personas suppress payoff-aligned behavior, with equilibrium attainment shifting by up to 90 percentage points depending on framing.
Large language models are increasingly deployed in multi-agent systems for strategic tasks, yet how design choices such as role-based personas and payoff visibility affect behavior remains poorly understood. We investigate whether LLM agents function as payoff-sensitive strategic actors or as identity-driven role followers. Using a 2x2 factorial experiment (persona presence x payoff visibility) with four models (Qwen-7B/32B, Llama-8B, Mistral-7B), we test 53 environmental policy scenarios in four-agent strategic games. We find that personas suppress payoff-aligned behavior: with personas present, all models achieve near-zero Nash equilibrium in Tragedy-dominant scenarios despite complete payoff information. Nearly every equilibrium reached is Green Transition. Removing personas and providing explicit payoffs are both near-necessary for payoff-aligned behavior, enabling only Qwen models to reach 65--90\% equilibrium rates. Our results reveal three behavioral profiles: Qwen adapts to framing, Mistral is disrupted without finding Tragedy equilibrium, and Llama remains near-invariant. We show that the same binary design choice can shift equilibrium attainment by up to 90 percentage points, establishing that representational choices are not implementation details but governance decisions.