Agent-Based Simulations of Online Political Discussions: A Case Study on Elections in Germany

2025-03-31

AI Summary

This study investigates how historical context, temporal constraints, and reward mechanisms shape users’ engagement in online political discourse on social media, using Twitter data from the German federal election as empirical basis. We propose the first history-aware multi-agent simulation framework, integrating fine-grained NLP models—namely, irony detection, offensiveness classification, and sentiment analysis—into a myopic best-response game to dynamically model user interactions under bounded rationality, cognitive resource limitations, and motivational constraints. Crucially, we introduce a novel mechanism for modeling historical dependence in AI-generated responses. Quantitative analysis reveals that prior dialogue history significantly biases generated content orientation, while time pressure and reward incentives exert nonlinear regulatory effects on participation dynamics. Our framework delivers an interpretable, scalable computational paradigm for analyzing political communication dynamics within algorithmically mediated environments.

Technology Category

• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Natural Language Processing: Discourse, Pragmatics & Argument Mining
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction

Application Category

• Social Networks and Social Media: Generative AI / large language models and their impact on social systems
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

User engagement on social media platforms is influenced by historical context, time constraints, and reward-driven interactions. This study presents an agent-based simulation approach that models user interactions, considering past conversation history, motivation, and resource constraints. Utilizing German Twitter data on political discourse, we fine-tune AI models to generate posts and replies, incorporating sentiment analysis, irony detection, and offensiveness classification. The simulation employs a myopic best-response model to govern agent behavior, accounting for decision-making based on expected rewards. Our results highlight the impact of historical context on AI-generated responses and demonstrate how engagement evolves under varying constraints.

Problem

• Modeling user interactions in online political discussions
• Analyzing impact of historical context on AI responses
• Studying engagement evolution under varying constraints

Innovation

• Agent-based simulation models user interactions
• AI models fine-tuned with sentiment analysis
• Myopic best-response governs agent behavior