Key points are not available for this paper at this time.
Research on political deepfakes has largely focused on detection accuracy, as well as broader psychological consequences such as uncertainty and mistrust. Less is known, however, about whether deepfake exposure influences issue-level political attitudes and under which conditions such effects occur. The present study investigated the attitudinal effects of political deepfakes on two polarizing issues, specifically climate change and immigration. In an online experiment conducted across Slovenia, Italy, and the United Kingdom ( N = 1124), participants were randomly exposed to one real video, a high-quality deepfake, or a low-quality deepfake promoting either positive or negative attitudes toward each topic. The study examined whether attitudes shifted after a single exposure and tested whether objective video characteristics, perceived trustworthiness, and political orientation moderated these effects. Deepfakes were about as persuasive as real videos, although overall attitudinal shifts were modest. Notably, perceived trustworthiness of the content emerged as a consistent moderator; deepfakes exerted stronger influence when their message was seen as credible. Perceived trustworthiness of the presentation (e.g., delivery, technical quality) also played a role, though more limited. Contrary to expectations, political orientation did not significantly moderate deepfake effects. The findings suggest that the persuasive power of deepfakes depends less on their technical sophistication and more on how credible their content appears to viewers, highlighting the need for interventions focused on fostering critical evaluation of political video content, particularly in the age of artificial intelligence-driven media manipulation.
Plohl et al. (Sun,) studied this question.