Political Bots and the Swedish General Election

Abstract

In this paper we present a study of political bots on Twitter and their influence on the Swedish general election in September 2018. We have not restricted our study to bots that are a software program, instead we are interested in any type of bot-like automated behaviour. This includes a human being manually copying or retweeting content repeatedly in a robot-like way with the aim of influencing the interaction between a user and content or with other users. We examined how bots influence discussions about the Swedish general election on Twitter during a three month period starting half a year before the election. For this purpose we trained a machine learning model to recognize automatic behaviour and studied accounts that tweet about the Swedish election. Our results show, among other things, that bot activity is at a similar level to that found in other studies, and that distinctive clusters, for example among right-wing Twitter accounts, can be identified.