Data Set.sav

The rise of deepfakes, AI-generated synthetic media, poses significant threats to the political arena. This study investigates how motivated reasoning influences responses to political deepfakes, and whether AI disclosure labels can encourage systematic information processing. In a mixed experiment with attitude consistency (consistent vs. inconsistent) and AI label (present vs. absent) as between-subject factors and time (pre vs. post) as a within-subject factor, 233 Republican participants were randomly exposed to AI-manipulated speeches of Donald Trump expressing either conservative or liberal immigration views, with or without AI labels. Results showed that attitude consistency significantly reduced skepticism toward deepfakes, while AI labels neither increased skepticism nor moderated the relationship between attitude consistency and skepticism. Lower skepticism led to greater attitude change on immigration policies and increased intentions to share deepfakes. Additionally, skepticism mediated the relationship between attitude consistency and sharing intentions. Theoretical contributions and implications are discussed.<br>