PRISM: A Personality-Driven Multi-Agent Framework for Social Media Simulation

2025-12-22

AI Summary

Conventional opinion dynamics models assume psychological homogeneity among individuals, limiting their ability to capture the interplay between cognitive biases and information diffusion in online polarization. Method: This paper proposes a multi-agent simulation framework tailored for social media, introducing—novelty—the first MBTI-type-driven cognitive strategy assignment mechanism. It integrates stochastic differential equations (SDEs) to model affective evolution, personality-conditioned partially observable Markov decision processes (PC-POMDPs) to formalize dynamic decision-making, and multimodal large language models (MLLMs) for social-data-driven agent initialization. Contribution/Results: Compared to homogeneous and Big-Five-based baselines, our framework significantly improves personality consistency and achieves high-fidelity reproduction of key emergent phenomena—including rational inhibition and affective resonance—thereby providing an interpretable, mechanism-grounded modeling approach to ideological polarization.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Knowledge Representation and Reasoning: Preferences

Application Category

• Social Networks and Social Media: Media and governance, opinion dynamics, filter bubbles, polarization
• User Modeling, Personalization and Recommendation: Psychology-informed user models and recommender systems
• Economics, Online Markets and Human Computation: Social networks and social learning

Abstract

Traditional agent-based models (ABMs) of opinion dynamics often fail to capture the psychological heterogeneity driving online polarization due to simplistic homogeneity assumptions. This limitation obscures the critical interplay between individual cognitive biases and information propagation, thereby hindering a mechanistic understanding of how ideological divides are amplified. To address this challenge, we introduce the Personality-Refracted Intelligent Simulation Model (PRISM), a hybrid framework coupling stochastic differential equations (SDE) for continuous emotional evolution with a personality-conditional partially observable Markov decision process (PC-POMDP) for discrete decision-making. In contrast to continuous trait approaches, PRISM assigns distinct Myers-Briggs Type Indicator (MBTI) based cognitive policies to multimodal large language model (MLLM) agents, initialized via data-driven priors from large-scale social media datasets. PRISM achieves superior personality consistency aligned with human ground truth, significantly outperforming standard homogeneous and Big Five benchmarks. This framework effectively replicates emergent phenomena such as rational suppression and affective resonance, offering a robust tool for analyzing complex social media ecosystems.

Problem

• Modeling psychological heterogeneity in online opinion dynamics
• Capturing interplay between cognitive biases and information propagation
• Simulating emergent social phenomena like rational suppression and affective resonance

Innovation

• Hybrid SDE and PC-POMDP framework for emotional and decision modeling
• MBTI-based cognitive policies for MLLM agents with data-driven priors
• Superior personality consistency replicating emergent social media phenomena