Decoding manipulative narratives in cognitive warfare: a case study of the Russia-Ukraine conflict

This study investigates the construction and dissemination of manipulative narratives in the context of cognitive warfare during the Russia-Ukraine conflict. Leveraging a mixed-methods approach that integrates AI-assisted semantic analysis with expert validation, we examine how adversarial messaging exploits cognitive biases-such as fear and confirmation bias-to influence perceptions and disrupt institutional trust.

Using the proprietary Attack-Index tool and large language models (LLMs), we detect linguistic markers of manipulation, including euphemisms, sarcasm, and strategic framing.

Our findings demonstrate that emotionally charged narratives, particularly those invoking nuclear threat scenarios, are synchronized with key geopolitical events to influence decision-makers and public opinion. The study identifies five thematic clusters and traces shifts in rhetorical strategies over time, showing how manipulative discourse adapts to geopolitical contexts. Special attention is given to the differentiated targeting of international political elites, Western publics, and Russian domestic audiences, each exhibiting varied cognitive vulnerabilities.

We acknowledge methodological and ethical limitations, including the dual-use potential of AI tools and challenges in establishing causal inferences. Nonetheless, this study offers the following key contributions:Empirically establishing nuclear rhetoric as a strategic element of narrative manipulation, particularly around NATO summits and military aid announcements.Advancing an integrated analytical framework that combines semantic clustering and AI-based discourse detection to monitor information threats in real time.Providing actionable insights for policy and digital security, including the development of countermeasures and international collaboration in addressing cognitive warfare.