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

TL;DR

PRISM addresses online polarization by modeling psychologically heterogeneous agents through MBTI-based policies and a hybrid dynamical core that links continuous emotional evolution via $d\mathbf{e}_i^{(t)}$ with discrete decision-making in a PC-POMDP. It grounds priors in a large-scale multimodal dataset and derives data-driven emotional parameters using $\hat{\theta}_{i,k} = P(e_k|\mathcal{T}_i)$ with Dirichlet smoothing and discretization $\epsilon$. The approach achieves superior alignment with human trajectories (e.g., $\rho=0.782$, $p<0.01$) and a notable reduction in polarity error (reported as $0.14$ MAE) while reproducing emergent phenomena such as rational suppression and affective resonance; ablation confirms the importance of individual MBTI dimensions. Overall, PRISM provides a robust, empirically grounded tool for mechanistically decoding complex social-media ecosystems and polarization dynamics using MBTI-informed agent policies and hybrid affective dynamics.

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.

Figures (4)

• Figure 1: Empirical Posterior Distributions of Affective States.
• Figure 2: The overall framework of PRISM.
• Figure 3: Change in average sentiment score of all agents over time during simulation of discussions on a controversial topic.
• Figure 4: Case study visualization. A multi-turn dialogue simulation of a thread.