Abstract Generative artificial intelligence (GenAI) is fundamentally changing political campaigning worldwide. Campaigns particularly rely on visual GenAI tools to create image and video material for strategic voter communication, rapidly producing large volumes of customized, high-quality content that is hardly distinguishable from ’real’ human creations. This has triggered public, media, and academic discussions about visual GenAI’s risks for voters, elections, and democracies, mainly centered on three claims: 1) visual GenAI will massively increase the quantity of campaign (dis)information, 2) it will raise the quality of such content through photorealistic visuals, and 3) AI-generated visuals will exert superior persuasive power over voters. However, empirical evidence tells a more differentiated story, indicating that the main challenges lie less in the technology itself than in how it is deployed. Based on current empirical studies, this article discusses how visual GenAI is used for political campaigning on social networking platforms, and how AI-generated visuals impact citizens’ political attitudes and behavior. These insights matter for the discussion of GenAI’s possibilities and challenges in political campaigning and especially for proposals on the responsible use of this technology.
Kruschinski et al. (Mon,) studied this question.