On the Dynamics of Multi-Agent LLM Communities Driven by Value Diversity

TL;DR

This paper investigates how diversity of human-like values influences the collective behaviors of multi-agent LLM communities, revealing that moderate diversity promotes stability and creativity, while extreme heterogeneity causes instability.

Contribution

It introduces a novel simulation framework grounded in Schwartz's Value Theory to study the impact of value diversity on AI community dynamics.

Findings

Value diversity enhances stability and emergent behaviors.

Moderate heterogeneity fosters creativity and innovation.

Extreme heterogeneity leads to instability.

Abstract

As Large Language Models (LLM) based multi-agent systems become increasingly prevalent, the collective behaviors, e.g., collective intelligence, of such artificial communities have drawn growing attention. This work aims to answer a fundamental question: How does diversity of values shape the collective behavior of AI communities? Using naturalistic value elicitation grounded in the prevalent Schwartz's Theory of Basic Human Values, we constructed multi-agent simulations where communities with varying numbers of agents engaged in open-ended interactions and constitution formation. The results show that value diversity enhances value stability, fosters emergent behaviors, and brings more creative principles developed by the agents themselves without external guidance. However, these effects also show diminishing returns: extreme heterogeneity induces instability. This work positions value diversity as a new axis of future AI capability, bridging AI ability and sociological studies of institutional emergence.