Manufacturing Consensus: Computational Propaganda and the 2016 U.S. Presidential Election

Abstract

This dissertation is an investigation of the ways three political actor groups used and interacted with bots and computational propaganda during the 2016 Presidential Election in the United States of America: political campaigns, journalists, and digital constituents. It is informed by data from over nine months of fieldwork, from February 2016 to after November 2016, in and around the U.S. campaign including attendance at pivotal party events, participant observation of campaigns, and consistent collection and parsing of online and offline information related to bots, computational propaganda, and the race. Over 40 interviews with a variety of experts from each of the three actor groups were done for this project. Three core working theoretical concepts emerged from this research: manufacturing consensus, the bot as an information radiator, and the bot as a proxy for the creator. The first describes the usage of bots and computational propaganda in attempts to amplify content online and give political ideas and actors the illusion of popularity in an effort to create bandwagon support. The second related to how journalists use bots as prostheses for reporting—to write simple stories, collect and parse data, and continuously communicate with the public about important information. The third explores how digital constituents, or citizens engaged in digital political communication and the usage of bots and computational propaganda online, and other bot builders can be theoretically conceived as related to, but separate from, the bots they create and deploy over the Internet. I argue, via each of these concepts and through the other findings of this research, that bots are one of the most important new tools for political communication in the United States. These automated software actors are also useful in understanding novel relationships between technology and society.