To Believe or Not to Believe: Framing Analysis of Content and Audience Response of Top 10 Deepfake Videos on YouTube

This study explores popular deepfake media content and audience response in an effort to gain better understanding of the potential social and psychological impacts of deepfakes. A content analysis of the top 10 YouTube deepfake videos and their audience comments (n = 2,689) was conducted to investigate the degree to which media meta-frame influenced audience response. The results suggested that media meta-frame, the type of video (deepfake videos with commentaries or original deepfake videos), and the number of dislikes on the video had considerable influence on audience response (attitudes and perceived realism), while the majority of the audience expressed neutral or irrelevant attitudes. Our exploratory observation revealed a dynamic interplay of how deepfake was presented and how others react to it might influence the public. Thus, it is necessary to appropriately caution the public about their vulnerability to the ever-advancing deepfake technologies.