Using LLMs to Advance the Cognitive Science of Collectives

2025-05-28

AI Summary

While large language models (LLMs) are widely applied in individual cognition research, their systematic use in collective cognition remains underexplored. Method: This paper pioneers the use of LLMs as a “computational sandbox” and “theoretical probe” for collective cognition, employing multi-agent simulation, prompt-driven group reasoning experiments, and a cognitive interpretability analysis framework to address methodological challenges arising from complex group interactions. We identify structural bias risks inherent in LLM-based simulations of group dynamics and systematically evaluate LLM capabilities—across concept generation, consensus evolution, and error cascade modeling—against empirically grounded cognitive benchmarks. Contribution/Results: The work establishes a novel, LLM-based paradigm for collective cognition research, offering a scalable, interpretable, and empirically verifiable methodology that bridges micro- and macro-cognitive scales in cognitive science.

Technology Category

• Cognitive Modeling & Cognitive Systems: Simulating Human Behavior
• Machine Learning: Large Multimodal Models (LMMs)
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

LLMs are already transforming the study of individual cognition, but their application to studying collective cognition has been underexplored. We lay out how LLMs may be able to address the complexity that has hindered the study of collectives and raise possible risks that warrant new methods.

Problem

• Applying LLMs to study collective cognition complexity
• Addressing underexplored collective cognition research gaps
• Identifying risks in LLM-based collective cognition studies

Innovation

• LLMs study collective cognition complexity
• LLMs address underexplored collective cognition
• New methods mitigate LLM risks