Can We Fix Social Media? Testing Prosocial Interventions using Generative Social Simulation

2025-08-05

AI Summary

This study investigates whether social media platforms can mitigate structural harms—such as political polarization and the decline of constructive dialogue—through prosocial interventions. We propose a generative social simulation framework that integrates large language models (LLMs) into agent-based modeling (ABM) to construct an interactive, high-fidelity virtual social platform. The model explicitly reproduces three key dysfunctions: echo chambers, elite-driven discourse, and polarization amplification. By simulating user posting, sharing, and following behaviors, we quantitatively analyze network evolution and information diffusion dynamics. Results indicate that mainstream intervention strategies yield limited efficacy—and some even exacerbate polarization—revealing that platform architecture (rather than algorithmic tuning alone) constitutes a fundamental causal mechanism underlying polarization. Our core contribution is the first instantiation of an LLM-ABM hybrid paradigm, providing a rigorously testable causal inference tool and actionable architectural redesign principles for platform governance.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Multiagent Systems: Agent-Based Simulation and Emergent Behavior

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Social media platforms have been widely linked to societal harms, including rising polarization and the erosion of constructive debate. Can these problems be mitigated through prosocial interventions? We address this question using a novel method - generative social simulation - that embeds Large Language Models within Agent-Based Models to create socially rich synthetic platforms. We create a minimal platform where agents can post, repost, and follow others. We find that the resulting following-networks reproduce three well-documented dysfunctions: (1) partisan echo chambers; (2) concentrated influence among a small elite; and (3) the amplification of polarized voices - creating a 'social media prism' that distorts political discourse. We test six proposed interventions, from chronological feeds to bridging algorithms, finding only modest improvements - and in some cases, worsened outcomes. These results suggest that core dysfunctions may be rooted in the feedback between reactive engagement and network growth, raising the possibility that meaningful reform will require rethinking the foundational dynamics of platform architecture.

Problem

• Mitigate social media harms like polarization and debate erosion
• Test prosocial interventions using generative social simulation
• Address core dysfunctions in platform architecture dynamics

Innovation

• Generative social simulation with LLMs
• Agent-Based Models for synthetic platforms
• Testing interventions like bridging algorithms