Abstract

This paper investigates how Instagram users make sense of the platform’s opaque AI-driven algorithmic recommendations. Building on Ytre-Arne and Moe’s (2020) folk-theory framework, we conduct 20 semi-structured interviews with active users from diverse backgrounds, genders, ages, and nationalities, followed by a thematic analysis to adapt and extend the framework to the context of Instagram. We corroborate and reconceptualize five established folk theories, confining, practical, reductive, intangible, and exploitative, within a social media environment, and identify three emergent theories distinctive to highly personalized social media platforms: controllable (users believe they can influence or manage their algorithm), representative (feeds are perceived as signaling interests and aspects of identity and are utilized as social cues to understand others), and influential (algorithms are understood as shaping users’ perceptions, opinions, and behaviors). 

These findings demonstrate that social media, particularly highly personalized platforms that utilize AI-driven algorithmic systems, is qualitatively different from media more broadly and that new folk theories are needed to understand how users make sense of algorithmic recommendations in such environments. By foregrounding newly identified folk theories, it offers a richer and more nuanced account of the interplay between users and algorithms in contemporary social media. Finally, the paper’s practical implications urge stakeholders to move beyond simplistic models of user-algorithm interaction toward more sophisticated, user-informed approaches to platform design, governance, and digital literacy, supporting healthier, more empowered, and more transparent digital environments.

Download

Abstract Only

Share

COinS