Multidimensional Mappings of Political Accounts for Malicious Political Socialbot Identification

Abstract

Malicious political socialbots used to sway public opinion regarding the U.S. government and its functions have been identified as part of a larger information warfare effort by the Russian government. This work asks what is knowable from a web-based sleuthing approach regarding the following four factors: 1) the ability to identify malicious political socialbot accounts based on their ego neighborhoods at 1, 1.5, and 2 degrees; 2) the ability to identify malicious political socialbot accounts based on the claimed and linked geographical locations of their accounts, their ego neighborhoods, and their #hashtag networks; 3) the ability to identify malicious political socialbot accounts based on their strategic messaging (content, sentiment, and language structures) on respective social media platforms; and 4) the ability to identify and describe “maliciousness” in malicious political socialbot accounts based on observable behaviors on that account on three social media platform types: (a) microblogging, (b) social networking, and (c) crowd-sourced encyclopedia content sharing.