Cognitive Cyber Crimes in the Era of Artificial Intelligence

3AI-Generated Propaganda and Automated Misinformation Warfare

Romil Rawat

LabGeoInf – Research LABoratory in GEOmatics and INFormation Systems, Rome, Italy

Abstract

The proliferation of artificial intelligence (AI) in information systems has led to an alarming rise in AI-generated propaganda and automated misinformation warfare. This study proposes a novel detection and analysis framework using the AI-MisInfoNet method, integrating semantic encoding, transformer-based attention, and cross-domain feature fusion. We used the FakeNewsNet dataset, which includes comprehensive data from PolitiFact and GossipCop domains encompassing news content, user engagements, and social context. Using a hybrid transformer-enhanced graph neural network, our model achieved 97.86% accuracy, 98.12% precision, and a false-positive rate of only 1.05%. Novel parameters such as Influence Spread Index and Narrative Coherence Drift were introduced to measure propaganda propagation and semantic deviation. The results demonstrate significant potential in mitigating automated misinformation threats through advanced AI models.

Keywords: AI-generated propaganda, misinformation warfare, FakeNewsNet, TE-GNN, semantic drift

3.1 Introduction

In recent years, the strategic misuse of artificial intelligence (AI) [1–3] to fabricate propaganda and conduct automated misinformation warfare [16–19] has intensified. Advanced generative models such as GPT-4, DALL·E 3 [4, 5], and stable diffusion are now capable of producing ...