From SERPs to Sound: How Search Engine Result Pages and AI-generated Podcasts Interact to Influence User Attitudes on Controversial Topics

TL;DR

This paper investigates how SERPs and AI-generated podcasts interact to influence user attitudes toward controversial topics. It uses a $2\times3\times2$ between-subjects design, manipulating information medium sequence, viewpoint bias, and topic controversiality, and analyzes data with a piecewise $Difference-in-Differences$ (DiD) mixed model. The findings show a significant sequence effect favoring podcast-first exposure, with nuanced interactions between bias and controversiality across the $\Delta_1$ and $\Delta_2$ segments, and no robust moderation by individual differences such as AI literacy or need for cognition. The work highlights practical implications for designing multimodal information ecosystems and responsible content delivery, and makes all data and code publicly available for reproducibility.

Abstract

Compared to search engine result pages (SERPs), AI-generated podcasts represent a relatively new and relatively more passive modality of information consumption, delivering narratives in a naturally engaging format. As these two media increasingly converge in everyday information-seeking behavior, it is essential to explore how their interaction influences user attitudes, particularly in contexts involving controversial, value-laden, and often debated topics. Addressing this need, we aim to understand how information mediums of present-day SERPs and AI-generated podcasts interact to shape the opinions of users. To this end, through a controlled user study (N=483), we investigated user attitudinal effects of consuming information via SERPs and AI-generated podcasts, focusing on how the sequence and modality of exposure shape user opinions. A majority of users in our study corresponded to attitude change outcomes, and we found an effect of sequence on attitude change. Our results further revealed a role of viewpoint bias and the degree of topic controversiality in shaping attitude change, although we found no effect of individual moderators.

Figures (3)

• Figure 1: Predicted attitude trajectories for each viewpoint bias across the three time points ($pre$, $mid$, $post$) for moderately controversial topics. Shaded regions represent 95% CIs.
• Figure 2: Predicted attitude trajectories for each viewpoint bias across the three time points ($pre$, $mid$, $post$) for highly controversial topics. Shaded regions represent 95% CIs.
• Figure 3: Predicted attitude trajectories for topics with varying controversiality (moderate, high) across the time points ($pre$, $mid$, $post$), separated by the medium sequence (SERP-first, Podcast-first), and with viewpoint bias fixed to neutral.