Evaluating LLM Agent Collusion in Double Auctions

TL;DR

This paper investigates how large language model agents behave in simulated double auction markets, revealing factors that influence their tendency to collude, with implications for economic stability and ethics.

Contribution

It systematically studies the emergence of collusion among LLM agents in market simulations, highlighting the effects of communication, model choice, and environmental pressures.

Findings

Communication increases collusion among agents.

Collusive behavior varies across different LLM models.

Environmental pressures can suppress or promote collusion.

Abstract

Large language models (LLMs) have demonstrated impressive capabilities as autonomous agents with rapidly expanding applications in various domains. As these agents increasingly engage in socioeconomic interactions, identifying their potential for undesirable behavior becomes essential. In this work, we examine scenarios where they can choose to collude, defined as secretive cooperation that harms another party. To systematically study this, we investigate the behavior of LLM agents acting as sellers in simulated continuous double auction markets. Through a series of controlled experiments, we analyze how parameters such as the ability to communicate, choice of model, and presence of environmental pressures affect the stability and emergence of seller collusion. We find that direct seller communication increases collusive tendencies, the propensity to collude varies across models, and environmental pressures, such as oversight and urgency from authority figures, influence collusive behavior. Our findings highlight important economic and ethical considerations for the deployment of LLM-based market agents.