The Application of Social Media in Modern-Day Influence Campaigns: Personality Profiling and Information Warfare

Abstract

Social media has become a repository of peoples’ information, where nearly 60% of the worlds population share ideas and exchange opinions. While this enables humans to be more connected than ever, it also creates an environment where peoples’ data can be used to manipulate opinions at large scales. Our work extends existing techniques which aim to quantify the extent to which public opinion and online discourse can be influenced by companies/governments. Firstly, we explore how an individuals online digital footprint can be used to understand personal attributes about them, such as their personality type. We then consider how this sort of information can be utilised for influence operations in modern-day conflicts, such as the 2022 Russia/Ukraine war. Personality profiling has been utilised by companies for targeted advertising, polit- ical campaigns and vaccine campaigns. However the accuracy and versatility of such models still remains relatively unknown. Consequently, we aim to explore the extent to which peoples’ online digital footprints can be used to profile their Myers-Briggs person- ality type. We analyse and compare the results of four models: logistic regression, naive Bayes, support vector machines and random forests. We discover that a support vector machine model achieves the best accuracy of 20.95% for predicting someones complete personality type. However, logistic regression models perform marginally worse and are significantly faster to train and predict, highlighting that relatively simple models can out- perform complex machine learning models. We acknowledge the presence of substantial class imbalance in our dataset and compare a number of methods for fixing the problems encountered with this. Moreover, we develop a statistical framework for assessing the importance of different sets of features in our models. We discover some features to be more informative than others in the Intuitive/Sensory (p = 0.032) and Thinking/Feeling (p = 0.019) models. While we apply these methods and models to Myers-Briggs person- ality profiling, they could be more generally used for any labelling of individuals on social media. The 2022 Russian invasion of Ukraine emphasises the role social media plays in modern-day conflicts, with both sides fighting in the physical and information environ- ments. There is a large body of work on identifying malicious cyber-activity, but less focusing on the effect this activity has on the overall conversation, especially with regards to the Russia/Ukraine Conflict. Here, we employ a variety of techniques including sentiment/linguistic analysis and time series analysis to understand how certain bot activity influences wider online discourse. In our results we observe that self declared bots most strongly increase discussions of work/governance (p = 3.803×10−18) with the most promi- nent effects after five hours. Moreover, we observe that self declared bots increase angst in the online discourse (p = 2.450 × 10−4) and discussions of motion (p = 7.93 × 10−10) with the most prominent effects after seven hours and three hours, respectively. Discussions of motion were most often involved with staying/fleeing a country and hence self-declared bots were likely influencing peoples’ decision to flee their country or not. Our work ex- tends and combines existing techniques to quantify how bots are influencing people in the online conversation around the Russia/Ukraine invasion. It provides a statistical frame- work which can be applied more generally to any influence campaign on social media and enables researchers to quantitatively understand what makes these campaigns impactful.