Beacken, G. D., Trauthig, I. K., & Woolley, S. C. (2026). Generative AI, propaganda, and digital authoritarianism: Comparative insights from six democratically weakened countries. Big Data & Society, 13. https://doi.org/10.1177/20539517251410064
Summary
This paper offers a comparative qualitative analysis of how generative AI (GenAI) is used for political propaganda in six democratically weakened countries: Bolivia, Hungary, India, Nigeria, Tunisia, and Turkey. Against prevailing narratives that treat GenAI’s political impact as inevitable, uniform, and transformative, the authors argue that adoption and use are strongly mediated by local geographic, political, economic, and social structures. The result is a heterogeneous landscape of GenAI-enabled propaganda that resists technologically deterministic framings and demands context-sensitive study within debates on digital authoritarianism and computational propaganda.
Key Contributions
- One of the first cross-national comparative studies of GenAI propaganda across multiple weakening democracies.
- A structurally grounded framework that foregrounds contextual variation over technological determinism.
- An extension of computational propaganda and digital authoritarianism scholarship to encompass GenAI-specific dynamics.
- Empirical grounding for calls to internationalize research on AI and political communication beyond Western democracies.
Methods
Comparative qualitative case analysis of six democratically backsliding countries — Bolivia, Hungary, India, Nigeria, Tunisia, and Turkey — examining how political elites in each context adopt and deploy generative AI for propaganda. The design privileges cross-case structural comparison over uniform technical measurement.
Findings
- Patterns of GenAI use for propaganda diverge markedly across the six cases rather than converging on a common template.
- Political, economic, social, and geographic conditions shape whether and how elites integrate GenAI into their propaganda repertoires.
- Assumptions of inevitable, transformative GenAI impact are not borne out; structural mediators matter more than the technology’s affordances alone.
- Digital authoritarianism increasingly incorporates GenAI, but in locally contingent forms.
Connections
This work speaks directly to Global South and non-Western studies of problematic AI-generated content, including Waight2026-ts, Waight2025-al, Gaw2025-ru, and Kansaon2025-id, which similarly resist universalizing accounts of GenAI harms. Its concern with elite political deployment of GenAI resonates with campaign- and propaganda-focused work such as Dodds2026-df and Schroeder2026-im, while its critique of technologically deterministic impact assumptions parallels effect-skeptical findings in Hackenburg2025-dj and Hackenburg2026-ud.