“I Don’t Think TikTok Really Cares About the Truth:” Experiences of Users Who Are Low Vision or Blind with Misinformation on TikTok

Abstract

Moderating misinformation on social media is a complex task of warning users about potentially harmful content while remaining reliable, unbiased, and non-judgmental. Though this is a valid concern, it doesn't exempt platforms like TikTok from making their soft moderation interventions inaccessible for users who are low vision or blind. Through interviews with 13 low vision or blind TikTok users, we learned that this was exactly the case - the informative cues used for soft moderation were inaccessible in 93% of the cases. To address this participatory exclusion, our participants proposed redesigns for navigable informative cues through auditory means or "audio frictions" that both warn the users and provide them with contextual information on why a particular content might be misleading, false, or generally harmful.