To further our past research on Russian bots, we developed a classifier that separates bots by political orientation using supervised machine learning. Our findings underline the fact that efforts to classify bots will always be two-fold: first, identifying the bots, then determining their orientation.
Stukal, Denis, Sergey Sanovich, Joshua A. Tucker, and Richard Bonneau. “For Whom the Bot Tolls: A Neural Networks Approach to Measuring Political Orientation of Twitter Bots in Russia.” SAGE Open 9, no. 2, (2019). https://doi.org/10.1177/2158244019827715
Apr 12, 2019
Area of Study
Computational propaganda and the use of automated accounts in social media have recently become the focus of public attention, with alleged Russian government activities abroad provoking particularly widespread interest. However, even in the Russian domestic context, where anecdotal evidence of state activity online goes back almost a decade, no public systematic attempt has been made to dissect the population of Russian social media bots by their political orientation. We address this gap by developing a deep neural network classifier that separates pro-regime, anti-regime, and neutral Russian Twitter bots. Our method relies on supervised machine learning and a new large set of labeled accounts, rather than externally obtained account affiliations or orientation of elites. We also illustrate the use of our method by applying it to bots operating in Russian political Twitter from 2015 to 2017 and show that both pro- and anti-Kremlin bots had a substantial presence on Twitter.
Public attention has focused on the Russian government and their alleged use of automatic social media accounts, or bots, to perform propaganda activities abroad. Perhaps because the scale of the operation in the Twittersphere is enormous, there has been no public attempt to dissect the population of Russian bots and determine their political orientation. Our previous research on the topic (in 2014-15) revealed that bots accounted for a surprisingly large proportion of Twitter activity. Specifically, we found that on most days more than half of the accounts tweeting in our collection of Russian-language tweets about politics were bots. Surprisingly, not all these bots were pro-regime.
To further this research and determine more about the political orientation of bots, we developed a classifier that separates pro-regime, anti-regime, and neutral Russian Twitter bots. We use supervised machine learning and a large set of labeled Russian political accounts from 2015 to 2017 on Twitter to show that both pro- and anti-Kremlin bots have a substantial presence on Twitter.
Our findings highlight the fact that any attempt to characterize the political activity of bots is always going to be a two-step process: methods to detect bots and the classification of bots by political orientation. Otherwise, we face the risk of attributing political motivations to bots that, at best, may be neutral and, at worst, may actually be supporting the opposite side in a political conflict.