Prompted Identity Degrades Cooperation in Multi-Agent LLM Systems

2026-09-28

AI Summary

This study addresses the phenomenon of "factionalism" in multi-agent large language model (LLM) systems, wherein exposing model identity labels severely impedes collaborative efficiency. This work provides the first formal definition and quantification of this effect, systematically evaluating collaborative performance under varying identity disclosure conditions through multi-agent game-theoretic experiments, reasoning benchmarks, and statistical analyses. The findings demonstrate that arbitrary labeling consistently induces fragmentation: labeled groups exhibit a 30% increase in decision rounds, a 55% rise in token consumption, and a decline in success rate from 96% to 81%. Accordingly, this paper proposes a mitigation strategy based on removing identity labels, which effectively eliminates anomalous behaviors and significantly enhances overall system performance.

Abstract

Multi-agent LLM systems increasingly mix models from several providers, yet exposing each agent's underlying model identity to its peers significantly impairs cooperation. We show that when agents are aware of each other's model family, the group splits into clusters, where agents prefer interacting with others carrying their same label, although nothing in the task rewards or asks for such a split. We argue that the label itself causes this split, which we define as $\textit{factionalism}$. We show and measure this phenomenon in two cooperative games and on a reasoning benchmark, with nine to twenty-five agents drawn from up to five open-weight model families. We further show that when the announced families are shuffled, or replaced by arbitrary labels, the factions still follow this information; when the label is removed, this behavior disappears. In strictly cooperative tasks, labeled groups spend on average $30\%$ more rounds and $55\%$ more tokens to reach a decision, and their success rate drops from $96\%$ to $81\%$. The effect replicates across tasks, group sizes and model families. Withholding identity labels from the agents is simple and effective mitigation.

Problem

• Multi-agent LLM systems
• Cooperation degradation
• Factionalism
• Model identity
• Prompted identity

Innovation

• Multi-Agent LLM Systems
• Factionalism
• Model Identity
• Cooperation Degradation
• Identity Withholding