Strategic Algorithmic Monoculture:Experimental Evidence from Coordination Games

2026-04-10

AI Summary

This study investigates how AI agents adjust behavioral similarity in response to incentives within multi-agent coordination settings, distinguishing between inherent similarity and strategic convergence. Through controlled coordination game experiments, it compares the behavioral modulation mechanisms of humans and large language models (LLMs). The work provides the first empirical differentiation and testing of “primary algorithmic monoculture” versus “strategic algorithmic monoculture,” revealing that while LLMs can modulate similarity in response to incentives, they exhibit higher baseline similarity than humans. Consequently, LLMs underperform humans in tasks requiring the maintenance of behavioral diversity, highlighting a current limitation in their capacity to preserve heterogeneity.

Technology Category

• Humans and AI: Learning Human Values and Preferences
• Multiagent Systems: Mechanism Design
• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

AI agents increasingly operate in multi-agent environments where outcomes depend on coordination. We distinguish primary algorithmic monoculture -- baseline action similarity -- from strategic algorithmic monoculture, whereby agents adjust similarity in response to incentives. We implement a simple experimental design that cleanly separates these forces, and deploy it on human and large language model (LLM) subjects. LLMs exhibit high levels of baseline similarity (primary monoculture) and, like humans, they regulate it in response to coordination incentives (strategic monoculture). While LLMs coordinate extremely well on similar actions, they lag behind humans in sustaining heterogeneity when divergence is rewarded.

Problem

• algorithmic monoculture
• coordination games
• multi-agent systems
• large language models
• behavioral heterogeneity

Innovation

• algorithmic monoculture
• strategic coordination
• large language models
• behavioral heterogeneity
• multi-agent systems