Computational analysis of Information Disorder in Cognitive Warfare

Abstract

Cognitive Warfare represents the modern evolution of traditional conflict, where the human mind emerges as the primary battleground, and information serves as a weapon to influence people’s thoughts, perceptions, and behaviors. Adopting the Information Disorder perspective, this work meticulously explores the phenomena associated with Cognitive Warfare, particularly as they spread across online social networks and media, to better understand their textual nature. In particular, the work focuses on specific cognitive weapons predominantly used by malicious actors in this context, such as the dissemination of misleading political news, junk science, and conspiracy theories. Therefore, the paper proposes an approach to identify, extract, and assess text-based features able to characterize the forms of Information Disorder involved in Cognitive Warfare. The proposed approach starts with a literature review and ends by assessing the identified and selected features through comprehensive experimentation based on a well-known dataset and conducted through the application of machine learning methods. In particular, by applying the Rough Set Theory and explainable AI it is found that features belonging to readability, psychological, and linguistic categories demonstrate a significant contribution in classifying the aforementioned forms of disorder. The obtained results are highly valuable as they can be leveraged to analyze critical aspects of Information Disorder, such as identifying the intent behind manipulated content and its targeted audience.