Exploring social media users’ experiences with algorithmic transparency cues

Abstract

All mainstream social media platforms now use algorithms to display recommended content, and some (e.g. Instagram, LinkedIn) have started showing what we call algorithmic transparency cues about why certain posts are recommended. However, little is known about what cues users see on their own feeds and how they experience them. Thus, using an online survey ( N = 515) of adult U.S. social media users, we gathered data about two research questions: (1) What types of algorithmic cues users find in their own feeds, and (2) their experiences with algorithms and their transparency. Content analysis of user-submitted screenshots and cue descriptions shows that most transparency cues refer to users’ behaviors, behaviors of others in their network, and sponsored posts. Furthermore, open-ended responses indicate that users have critical opinions about algorithms, calling for greater algorithmic transparency on social media, and offering suggestions for researchers and platform designers moving forward.