A User Study on the Feasibility of Topic-aware Misinformation Warning on Social Media
Event Type
TimeWednesday, October 12th8:20am - 8:40am EDT
DescriptionMisinformation is one of the most fundamental problems in social media with increasing cases and underlying harmful effects on users. To mitigate such problem, misinformation warnings have been developed, including alerting with warning messages and hiding the contents. Previous studies mainly explored the most effective, one-size-fits-all design. Therefore, little has been known about customizable and flexible warning designs. In this study, we propose a "topic-aware misinformation warning" where users' preferences for warning designs can vary on topics. To illustrate our ideas, we developed Twitter-like pages using three topics (i.e., politics, gossip, and Covid-19) and three designs (i.e., interstitial, contextual, and highlight). We conducted semi-structured interviews with 18 participants to explore their preferences and opinions on the designs. Our results show that users' preferences for misinformation warnings are diverse in topics. Thus, topic-aware misinformation warning is promising to alleviate misinformation problems on Twitter.