Campaign Deepfakes and Affective Polarization: The Role of Artificial Intelligence in Campaigns in Shaping Voter Attitudes

Abstract

Does exposure to news about out-party campaigns’ increasing use of deepfakes further increase polarization among partisans, and how does this impact vary among different evaluations of political coalitions, parties, and leaders? What role does approval for candidates’ usage of AI play as a moderator in this relationship? This paper examines these questions using data from a survey experiment (N = 1627), in which each respondent exposed to a news story that varied in their coverage of the growing deployment of AI-powered deepfakes during the 2024 India’s national election campaigns. The findings indicate that exposure to news about out-party campaigns’ use of deepfakes did not polarize attitudes towards coalitions, parties, and leaders. In contrast, approval for candidates’ use of AI was negatively associated with polarization towards coalitions, parties, and leaders, but this did not moderate the experimental effects: respondents who approved of candidates’ AI usage and saw a news article about out-party use of deepfakes were not more polarized toward either coalitions, parties or leaders. This study not only extends the understanding of AI’s role in campaigns by examining its complex effects on affective polarization, but it also broadens empirical research into the impact of deepfakes by focusing on the non-Western context of India, thereby offering new theoretical insights into how technological advancements intersect with political behaviors and attitudes.