AI Agents Alone Are Not (Yet) Sufficient for Social Simulation

TL;DR

This paper argues that current LLM-based agents are insufficient for social simulation due to mismatches between their capabilities and the scientific requirements, emphasizing the need for explicit mechanisms and evaluation methods.

Contribution

It introduces a unified Markov game framework for AI agent-based social simulation, highlighting mechanisms for better design, evaluation, and interpretation.

Findings

LLM agents lack human behavioral validity in social simulations.

Simulation outcomes are often influenced by environment and scheduling, not just agent messaging.

Proposes explicit exposure and scheduling mechanisms for more auditable simulations.

Abstract

Recent advances in large language models (LLMs) have spurred growing interest in using LLM-integrated agents for social simulation, often under the implicit assumption that realistic population dynamics will emerge once role-specified agents are placed in a networked multi-agent setting. This position paper argues that LLM-based agents alone are not (yet) sufficient for social simulation. We attribute this over-optimism to a systematic mismatch between what current agent pipelines are typically optimized and validated to produce and what simulation-as-science requires. Concretely, role-playing plausibility does not imply faithful human behavioral validity; collective outcomes are frequently mediated by agent-environment co-dynamics rather than agent-agent messaging alone; and results can be dominated by interaction protocols, scheduling, and initial information priors. To make these underlying mechanisms explicit and auditable, we propose a unified formulation of AI agent-based social simulation as an environment-involved Markov game with explicit exposure and scheduling mechanisms, from which we derive concrete actions for design, evaluation, and interpretation.