``Learn the Facts About {COVID-19}'': {A}nalyzing the Use of Warning Labels on {TikTok} Videos

During the COVID-19 pandemic, health-related misinformation and harmfulcontent shared online had a significant adverse effect on society. To mitigatethis adverse effect, mainstream social media platforms employed soft moderationinterventions (i.e., warning labels) on potentially harmful posts. Despite therecent popularity of these moderation interventions, we lack empirical analysesaiming to uncover how these warning labels are used in the wild, particularlyduring challenging times like the COVID-19 pandemic. In this work, we analyzethe use of warning labels on TikTok, focusing on COVID-19 videos. First, weconstruct a set of 26 COVID-19 related hashtags, then we collect 41K videosthat include those hashtags in their description. Second, we perform aquantitative analysis on the entire dataset to understand the use of warninglabels on TikTok. Then, we perform an in-depth qualitative study, usingthematic analysis, on 222 COVID-19 related videos to assess the content and theconnection between the content and the warning labels. Our analysis shows thatTikTok broadly applies warning labels on TikTok videos, likely based onhashtags included in the description. More worrying is the addition of COVID-19warning labels on videos where their actual content is not related to COVID-19(23% of the cases in a sample of 143 English videos that are not related toCOVID-19). Finally, our qualitative analysis on a sample of 222 videos showsthat 7.7% of the videos share misinformation/harmful content and do not includewarning labels, 37.3% share benign information and include warning labels, andthat 35% of the videos that share misinformation/harmful content (and need awarning label) are made for fun. Our study demonstrates the need to developmore accurate and precise soft moderation systems, especially on a platformlike TikTok that is extremely popular among people of younger age.