On the Principles behind Opinion Dynamics in Multi-Agent Systems of Large Language Models

TL;DR

This paper investigates how opinions evolve among large language models in a simulated environment, revealing biases, the impact of interaction, and the effects of opinion formation methods on diversity and consensus.

Contribution

It introduces a novel analysis of opinion dynamics in LLM populations, highlighting biases, the influence of interaction, and the role of opinion formation processes.

Findings

Biases influence opinion exchange and consensus formation.

Diverse final opinions emerge when LLMs freely form opinions.

Consensus is more common when opinions are limited to multiple-choice options.

Abstract

We study the evolution of opinions inside a population of interacting large language models (LLMs). Every LLM needs to decide how much funding to allocate to an item with three initial possibilities: full, partial, or no funding. We identify biases that drive the exchange of opinions based on the LLM's tendency to find consensus with the other LLM's opinion, display caution when specifying funding, and consider ethical concerns in its opinion. We find these biases are affected by the perceived absence of compelling reasons for opinion change, the perceived willingness to engage in discussion, and the distribution of allocation values. Moreover, tensions among biases can lead to the survival of funding for items with negative connotations. We also find that the final distribution of full, partial, and no funding opinions is more diverse when an LLM freely forms its opinion after an interaction than when its opinion is a multiple-choice selection among the three allocation options. In the latter case, consensus is mostly attained. When agents are aware of past opinions, they seek to maintain consistency with them, changing the opinion dynamics. Our study is performed using Llama 3 and Mistral LLMs.