Key points are not available for this paper at this time.
Political deepfakes are regarded as an urgent threat to democracy. However, we know little about how different degrees of deception in deepfakes affect recipients’ credibility perceptions and the delegitimization of the target of the deepfake. Against this backdrop, we conducted a comparative experiment in the US and the Netherlands (N = 1187) to map the effects of domestic and foreign political deepfakes delegitimizing a prominent Democrat politician. Key findings illustrate that a hyper-realistic deepfake with plausible manipulation of content is perceived as substantially more credible than deepfakes with implausible content manipulation. However, implausible deepfakes have a stronger effect on delegitimizing the depicted politician. We also show that higher scores on a cognitive reflection test correspond to a lower credibility rating for deepfakes. Our findings demonstrate that political deepfakes do not have to be rated as credible to delegitimize established politicians and suggest that analytical thinking may enhance resilience toward deepfakes.
Hameleers et al. (Thu,) studied this question.