How building social media personas shapes political polarization
From Caricature to Collective Artifact: How Persona Construction Drives Polarization in Generative Social Simulations
Social and Information NetworksArtificial Intelligence
Summary
This paper looks at how creating online characters, or personas, with political views can lead to more division between groups when they interact on social media. The authors show that small changes in how these personas express their political opinions can greatly change how much people connect across parties. They also explain that treating every survey answer as equally important in persona creation can make these virtual characters act in overly extreme or simplified ways. The paper suggests ways to better evaluate such personas in future studies.
What this means in practice
- •For social media platform designers: Design simulations to test how political personas influence user interactions and reduce polarization by adjusting how strongly partisan traits are encoded.
- •For political campaign strategists: Use refined persona models to better understand the effects of political messaging and voter behavior in online social contexts.
Authors
Dimitris Michailidis, Petter Törnberg, Fernando P. Santos
Abstract
Generative Social Simulation (GSS) leverages Large Language Models (LLMs) to create sophisticated personas that define agents with demographics, emotions, and political opinions. While prior work shows that personas compress outputs into stereotypical caricatures, there is no comprehensive study of how these caricatures evolve into collective artifacts when they interact. Using a Twitter-like GSS platform with personas derived from the American National Elections Study (ANES), we ablate three layers of partisan identity: affective polarization markers, party identification, and vote choice, while retaining demographics, ideological positioning and important issues. We find that collective outcomes depend heavily on how survey responses are translated into personas. Removing a few explicitly partisan statements moves the network from near-complete segregation to cross-party ties approaching those of real Twitter networks. We argue that matching one outcome is not evidence for behavioral validity and point to salience as a possible explanation. Surveys ask for a position on every question regardless of its importance to the respondent, and persona construction turns each answer into a direct behavioral instruction. We conclude with a checklist for more comprehensive evaluation of personas in future works.