DISCURSIVE SECURITY IN THE AGE OF GENERATIVE ARTIFICIAL INTELLIGENCE: THREATS, VULNERABILITIES AND COMMUNICATIVE RESILIENCE

Abstract

The article examines discursive security as a specific dimension of information security in the age of generative artificial intelligence. It argues that AI does not merely increase the scale of information production but transforms the communicative conditions under which credibility, authority, expertise and public trust are constructed. The central problem is not limited to automated disinformation – it concerns the capacity of generative systems to simulate institutional language, manufacture synthetic credibility, personalise influence and intensify manipulative narratives. The study is theoretical and conceptual rather than empirical. It uses conceptual reconstruction and a critical synthesis of security studies, disinformation research, discourse analysis and selected legal sources. It proposes a model based on the triad of threats, vulnerabilities and resilience. Threats include AI-assisted propaganda, deepfakes, phishing, institutional impersonation, synthetic public opinion, and simulated expertise, while vulnerabilities include low media literacy, polarisation, emotionalised reception, weak source verification, and excessive trust in fluent, expert-like language. Communicative resilience requires more than technical detection and fact-checking; it involves critical discourse awareness, transparent institutional communication, source-verification practices, media and AI literacy, and proportionate legal safeguards. The article concludes that discourse should be treated as a central arena of contemporary security, since generative AI increasingly operates through the management of perception, credibility and trust.