Deepfake Architectures, Automated Synthetic Propaganda, and the Erosion of Institutional Trust in Post-Conflict Elections

Global Journal of Peace and Conflict Studies | Vol. 15, No. 1, January 2024 | pp. 1–9

DOI: 10.46882/2024/GJPCS/150101

Article Type: Original Research Paper

Title: Deepfake Architectures, Automated Synthetic Propaganda, and the Erosion of Institutional Trust in Post-Conflict Elections

Authors: Sarah L. Bernstein¹, Amanuel T. Mengistu²

Authors’ Affiliations: ¹Oxford Internet Institute, University of Oxford, Oxford, United Kingdom; ²Institute for Peace and Security Studies, Addis Ababa University, Addis Ababa, Ethiopia

Abstract:
The rapid global deployment of generative deepfake model architectures and highly automated algorithmic synthetic propaganda networks has structurally altered contemporary political discourse and accelerated identity-based polarization in fragile states. This paper investigates the digital mechanisms through which synthetic audiovisual campaigns manipulate public sentiment and amplify ethno-political divisions during electoral cycles. Employing a cross-national quantitative content analysis of approximately 1,150,000 localized political posts, we track the spread of automated, micro-targeted alternative narratives across digital spaces during recent constitutional transitions. The empirical results demonstrate that social media algorithms engineered to maximize user engagement disproportionately prioritize emotionally volatile, AI-generated multimedia content, creating isolated ideological echo chambers. Statistical correlation modeling establishes that a 20% increase in localized network exposure to verifiably false, deep-fake political media precedes spikes in urban civil unrest by a predictable 24-hour window (r = 0.71, p < 0.01). The study highlights how traditional civil society organizations are systematically marginalized within these digital architecture frameworks. We conclude that mitigating digital radicalization requires shifting from reactive content moderation toward imposing legally binding transparency requirements on algorithmic design models.

Keywords: Deepfake Architecture; Synthetic Propaganda; Algorithmic Polarization; Digital Warfare; Civil Discourse; Democratic Trust

Manuscript Timeline: Received: November 15, 2021; Revised: January 18, 2022; Accepted: March 02, 2022; Published: January 12, 2024

Citation: Bernstein, S. L., & Mengistu, A. T. (2024). Deepfake Architectures, Automated Synthetic Propaganda, and the Erosion of Institutional Trust in Post-Conflict Elections. Global Journal of Peace and Conflict Studies, 15(1), 1–9. DOI: 10.46882/2024/GJPCS/150101

Subscribe to read the full article: https://internationalscholarsjournals.org/subscribe-to-read