Title: Characterizing Retweet Bots: The Case of Black Market Accounts

Abstract: Malicious Twitter bots are detrimental to public discourse on social media. Past studies have looked at spammers, fake followers, and astroturfing bots, but retweet bots, which artificially inflate content, are not well understood. In this study, we characterize retweet bots that have been uncovered by purchasing retweets from the black market. We detect whether they are fake or genuine accounts involved in inauthentic activities and what they do in order to appear legitimate. We also analyze their differences from human-controlled accounts. From our findings on the nature and life-cycle of retweet bots, we also point out several inconsistencies between the retweet bots used in this work and bots studied in prior works. Our findings challenge some of the fundamental assumptions related to bots and in particular how to detect them.