Title:Characterizing and Detecting State-Sponsored Troll Activity on Social Media

Abstract:The detection of state-sponsored trolls acting in information operations is an unsolved and critical challenge for the research community, with repercussions that go beyond the online realm. In this paper, we propose a novel AI-based solution for the detection of state-sponsored troll accounts, which consists of two steps. The first step aims at classifying trajectories of accounts' online activities as belonging to either a state-sponsored troll or to an organic user account. In the second step, we exploit the classified trajectories to compute a metric, namely "troll score", which allows us to quantify the extent to which an account behaves like a state-sponsored troll. As a study case, we consider the troll accounts involved in the Russian interference campaign during the 2016 US Presidential election, identified as Russian trolls by the US Congress. Experimental results show that our approach identifies accounts' trajectories with an AUC close to 99% and, accordingly, classify Russian trolls and organic users with an AUC of 90%. Finally, we evaluate whether the proposed solution can be generalized to different contexts (e.g., discussions about Covid-19) and generic misbehaving users, showing promising results that will be further expanded in our future endeavors.