Generative Artificial Intelligence and Far-right Visual Communication on Social Media: Affordances, Actors, and Audiences

• In Event: Digital (Un)Democracy: AI, Media, and the New Political Order

Abstract

Generative AI tools are increasingly used to create synthetic images and videos for social media, and right-wing actors have been early adopters, producing visual propaganda ranging from images posted by activists to videos shared by Donald Trump on TruthSocial. This paper investigates how far-right activists, politicians, and their audiences have adopted generative AI to produce and circulate political imagery. I analyze the relationships between technological affordances, political actors, and online audiences, situating the production and dissemination of political AI-generated imagery within online information ecosystems characterized by polarization, fragmentation, and algorithmic amplification. I conduct three case studies, drawing upon primary documents collected from social media and AI company websites between August 2024 and February 2026, alongside extensive secondary media coverage. The paper explores how prominent British far-right activists have incorporated AI into their digital repertoires; how Trump and his administration have used AI imagery and how online audiences use these tools to co-produce participatory propaganda; and how Grok’s integration into X has opened new affordances for grassroots content creation. I demonstrate how these tools have enabled actors to flood online spaces with propaganda images and videos, augmenting rather than replacing existing forms of content. Much of the AI-generated political content takes the form of memes and stylized imagery rather than realistic deepfakes. However, improvements in generative image editing have opened new avenues for manipulating existing imagery, which political actors are already exploiting. Audiences actively use these same tools to amplify and contest elite communications, facilitating new forms of participatory computational propaganda. I conclude by emphasizing how generative AI has become an increasingly contested political object, as AI companies have drawn criticism for the kinds of material their models generate or refuse.

Author

• Thomas Davidson, Rutgers University-New Brunswick