Recognizing, Experiencing, and Navigating Eating Disorder Echo Chambers on TikTok: A Qualitative Study

Echo chambers are algorithmically curated environments characterized by reduced informational diversity and repeated exposure to content that reinforces existing attitudes and behaviors. Emerging research suggests TikTok's recommendation algorithm amplifies eating disorder (ED)-related content among individuals with EDs, creating ED echo chambers.

This study explored how individuals reporting current EDs recognize, experience, and navigate algorithmically curated ED echo chambers on TikTok.

Twenty-five TikTok users self-reporting current EDs completed semi-structured Zoom interviews exploring perceptions of TikTok's algorithm and experiences of navigating and disengaging from recommended ED-related content. Data were analyzed using reflexive thematic analysis.

Three themes were identified. The first (Becoming algorithmically enclosed) captured participants' recognition of ED echo chambers and two pathways into them: gradual escalation from health and wellness content and deliberate engagement with ED-related content. The second (Cognitive, emotional, and social experiences within ED echo chambers) described how upward comparison, validation, and belonging reinforced ED-related cognitions, behaviors, and identities. The third (Remaining enclosed) revealed that ED echo chambers were sustained by psychological mechanisms (emotional dependency, ED-centred identity maintenance, distress as toxic motivation) and structural mechanisms (ease of acquiring but difficulty removing ED-related content, ambiguity of ED-related messaging, paid advertisements). Recovery-oriented content failed to help because it remained embedded within the same ecosystem and maintained an ED-centred identity. Only a minority described partially exiting ED echo chambers after years of retraining their algorithm.

Participants described ED echo chambers as reinforcing ED pathology and resisting escape, suggesting individual self-regulation is insufficient and that clinically informed platform-level interventions are needed.