Graph Model of Information-Psychological Warfare

• Date: 2025
• Main Authors: Качинський, А. Б., Ланде, Д. В.
• Format: Article
• Language: Ukrainian
• Published: Інститут проблем реєстрації інформації НАН України 2025
• Subjects: mental wars; hierarchical model; generative artificial intelligence; large language models; semantic network; clustering; modularity
• Online Access: http://drsp.ipri.kiev.ua/article/view/335791
• Journal Title: Data Recording, Storage & Processing

This study presents a comprehensive graph-based model of information-psychological mental wars, offering a structured approach to understanding the complex mechanisms of influence in hybrid warfare. Building upon an initial hierarchical framework, the authors enhance and expand the model using generative artificial intelligence (AI) and large language models (LLMs), enabling the identification of new concepts, clusters, and interconnections that were previously overlooked. The model is structured around five core levels: goals of mental warfare, forces and means, actors, actors' objectives, and implementation policies. These elements are formalized into a directed graph where nodes represent concepts and edges illustrate functional dependencies, forming a semantic network capable of dynamic evolution. By integrating AI-generated insights with expert analysis, the model transitions from a rigid hierarchy to a flexible network structure, allowing for more nuanced and adaptive representations of mental warfare systems. This transformation enables the discovery of non-linear pathways toward strategic outcomes and facilitates more efficient resource allocation. The methodology includes iterative refinement through virtual expert collaboration, linguistic data processing, and visualization tools such as Gephi. The study applies clustering based on modularity classes and node ranking via algorithms like PageRank to identify key influential concepts within the network. Among the most impactful elements identified are disinformation, media manipulation, religion, culture, identity erosion, and economic influence. These findings provide critical insight into the objectives, strategies, and consequences of modern mental warfare, particularly in the context of Russia’s prolonged war against Ukraine. Ultimately, the extended network model enhances analytical capabilities for assessing hybrid threats and offers a robust framework for predicting and countering psychological and informational influence operations. The methodology presented can be applied across various domains requiring deep structural analysis of complex socio-political phenomena. Tabl.: 1. Fig.: 1. Refs: 13 titles.