Making up Audience: Media Bots and the Falsification of the Public Sphere

Abstract

ABSTRACT The purpose of this paper is to discuss if and how Brazilian media outlets make use of automated strategies and artificial intelligence (AI) in order to produce convenient social media metrics about themselves and amplify their relevance on Twitter. We examine how media bots can manipulate online ratings, change social perception of what is relevant and increase engagement with both on- and offline media entities. We extracted three types of data: (i) 530,942 tweets containing at least one URL from Globo Group or Folha Group collected via Twitter API; (ii) URL metadata from 158,690 articles by Globo Group and Folha Group; (iii) Twitter trending topics in Brazil. Profiles that posted links were later sampled and classified using the Botometer. Automated and human accounts were analyzed regarding their posting frequency and speed. In this paper, we assess the hypothesis that the existence of media bots is affecting the Twittersphere in Brazil, where automated accounts, empowered by AI, might be responsible for a substantial share of the links to popular Brazilian media outlets on Twitter. Our research provides quantitative empirical evidence that bots are particularly active in amplifying news media links in the initial moments of spreading. Additionally, automated accounts play an important role in promoting TV broadcast programs in Brazil. Based on these evidences, we discuss the strategies adopted by Brazilian media corporations to sustain their omnipresence online that boosts their online audience.