Analyzing the Behaviour of Twitter Bots in Post Brexit Politics

Abstract

Twitter bots are automated user accounts widely used in political campaigns to promote opinions and attack opponents. In this paper, we study the behaviour of bots in the Brexit debates after the 2016 referendum, including on issues such as a second Brexit referendum and Scottish independence. Our findings, which are useful for understanding the role of bots in political debates more generally, encourage further research on the effects of bots on public opinion. We collected our data over a set of Brexit related hashtags and searched for bot accounts by selecting strongly opinionated accounts, then analysed their strategies, intended influence and policy position on Brexit issues. There are more than 1,962 bot accounts currently engaged in the Brexit debate. Using a novel approach for reverse engineering twitter bots [5] we show how to uncover the bots' implemented strategies. We found that@StillYesScot,@IsThisABOt and@FAO_Scotbot, all bots that promote Scottish independence, use similar strategies. Such an analysis can contribute to identifying their political background and affiliation. Our differential sentiment analysis shows how the bots attempt to influence, constructively or destructively, the opinions of users connected to their network. Our geolocation analysis shows that the bots propagate their information to the correct target audience, with the majority of their followers living in the UK.