The Search for Balance: The Impact of LLM-based Conversational Search on Information Processing and Polarization

The emergence of large language models (LLMs) has fundamentally transformed how individuals access and interpret information, potentially shaping their opinions and attitudes. Our study investigates the impact of LLM-based conversational search on polarization, a significant concern in today’s digital landscape. Building upon information processing theory, we propose that LLM-based conversational search (with and without references) will prompt heuristic information processing compared to traditional search engines, thereby reducing opinion polarization. Furthermore, we explore the moderating effect of intensity of prior attitudes on the relationship between information search tools and polarization. A laboratory experiment is proposed to empirically test the theoretical model, and experimental platforms have already been implemented. This short paper aims to contribute to the literature on information search, polarization, and LLMs, while offering practical insights for the design of LLM interactions and governance.