Do discussions in human-computer communities trigger group polarization? Insights from the media evocation paradigm

Abstract

This study investigates group polarization and the spiral of silence within human-computer communities, particularly as intelligent chatbots become increasingly integrated into online interactions. Grounded in the media evocation paradigm, two experiments were conducted: Experiment 1, using a 4 × 4 design, explored how varying stances held by humans and chatbots influence group polarization, revealing distinct polarization mechanisms based on the differing stances within online communities. Experiment 2 employed a 3 × 4 design to examine the impact of human identifiability on the spiral of silence, finding that higher identifiability led to increased conformity to majority opinions, while full anonymity intensified the spiral of silence, especially when chatbots held strong stances. These results contribute to the understanding of group polarization, the media evocation paradigm, and the spiral of silence within human-computer communities.