Imani N. Sherman (University of Florida), Jasmine D. Bowers (University of Florida), Keith McNamara Jr. (University of Florida), Juan E. Gilbert (University of Florida), Jaime Ruiz (University of Florida), Patrick Traynor (University of Florida)
Robocalls are inundating phone users. These automated calls allow for attackers to reach massive audiences with scams ranging from credential hijacking to unnecessary IT support in a largely untraceable fashion. In response, many applications have been developed to alert mobile phone users of incoming robocalls. However, how well these applications communicate risk with their users is not well understood. In this paper, we identify common real-time security indicators used in the most popular anti-robocall applications. Using focus groups and user testing, we first identify which of these indicators most effectively alert users of danger. We then demonstrate that the most powerful indicators can reduce the likelihood that users will answer such calls by as much as 43%. Unfortunately, our evaluation also shows that attackers can eliminate the gains provided by such indicators using a small amount of target-specific information (e.g., a known phone number). In so doing, we demonstrate that anti-robocall indicators could benefit from significantly increased attention from the research community.