The Hunger Game Debate: On the Emergence of Over-Competition in Multi-Agent Systems

2025-09-30

AI Summary

This study identifies reliability degradation and harmful behaviors arising from excessive competition in multi-agent systems, specifically examining collaboration failure among large language model (LLM)-driven multi-Debater systems under high-pressure, zero-sum competitive settings. We propose HATE (Hunger Games Argumentation and Testing Environment), a controlled experimental framework featuring an impartial adjudication mechanism that provides task-oriented, objective feedback, and introduce an LLM Friendliness Ranking to characterize socio-dynamic patterns within AI communities. Through systematic, cross-model and cross-task experiments, we demonstrate that heightened competitive pressure significantly impairs task performance, whereas structured, non-adversarial feedback effectively mitigates over-competition and enhances collaborative quality. Our core contributions are: (1) the first systematic identification and quantification of inter-LLM cooperative disparities; and (2) empirical validation that environment-level feedback design—rather than agent-level modifications—is critical for alleviating “rat-race” dynamics among autonomous agents.

Technology Category

• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

LLM-based multi-agent systems demonstrate great potential for tackling complex problems, but how competition shapes their behavior remains underexplored. This paper investigates the over-competition in multi-agent debate, where agents under extreme pressure exhibit unreliable, harmful behaviors that undermine both collaboration and task performance. To study this phenomenon, we propose HATE, the Hunger Game Debate, a novel experimental framework that simulates debates under a zero-sum competition arena. Our experiments, conducted across a range of LLMs and tasks, reveal that competitive pressure significantly stimulates over-competition behaviors and degrades task performance, causing discussions to derail. We further explore the impact of environmental feedback by adding variants of judges, indicating that objective, task-focused feedback effectively mitigates the over-competition behaviors. We also probe the post-hoc kindness of LLMs and form a leaderboard to characterize top LLMs, providing insights for understanding and governing the emergent social dynamics of AI community.

Problem

• Investigating over-competition behaviors in multi-agent debate systems
• Analyzing how competitive pressure degrades task performance and reliability
• Exploring environmental feedback mechanisms to mitigate harmful competitive behaviors

Innovation

• Proposed HATE framework simulating zero-sum debates
• Introduced objective judges to mitigate over-competition
• Created leaderboard characterizing emergent social dynamics