Disordered Information: Commercial Care and Algorithmic Recoveries in Youtube "Eating Disorder Stories"

Abstract (English)

YouTube is rife with 'storytime' vlogs: informal, minimally-edited, conversational videos in which the creator tells their audience about events taking place in their daily life. A unique sub-branch of this video genre is 'eating disorder stories', in which creators often share graphic accounts of their battles with eating disorders from illness onset to debilitating symptoms, treatment, and eventual recovery. In a context where eating disorder prevalence continues to grow (Statistics Canada, 2022), and social scientists identify a growing trend toward digital patient-to-patient health information exchange (Philips and Rees, 2017; Kent, 2024), social media platforms like YouTube become critical resources to understand the kinds of health information people are engaging with. The fact that some of the most popular platforms, including YouTube, Instagram, and TikTok, are driven by algorithmic recommendation systems also presents a crucial opportunity for ethnographic exploration as algorithms profile viewers and suggest the content they see. Drawing on the algorithmic recommendations made to a YouTube profile tailored to represent a 'typical' person with an eating disorder, this paper explores the content and discourses shared within YouTuber ‘eating disorder stories’ from the perspective of a struggling viewer. It examines the kinds of information viewers are receiving, and discusses the implications of this form of information-sharing given the entanglement of commercial interests and algorithmic recommendations in content creation. It argues that conflicting interests between financial and algorithmic success on the one hand, and genuine care on the other, ultimately lead to the sharing of information and ideas about eating disorders that stand to harm viewers far more than they help. To illustrate the ways algorithm-driven social media affects health experiences, this paper also offers the term “algorithmic recoveries” both as a useful descriptor and as an invitation for future research.