Master’s Thesis Defense Meeting
Date: Friday, December 11th, 2020
Time: 1:00pm
Location: Virtual, https://bluejeans.com/256884510
 
Advisor:
Richard Catrambone, Ph.D. (Georgia Tech)
 
Thesis Committee Members:
Richard Catrambone, Ph.D. (Georgia Tech)
Bruce N. Walker, Ph.D. (Georgia Tech)
Sidni Justus, Ph.D. (Oglethorpe)
 
Title: Individual Differences in Deepfake Detection: Mindblindness and Political Orientation

 Abstract: The proliferation of the capability for producing and distributing deepfake videos threatens the integrity of systems of justice, democratic processes, and the general ability to critically assess evidence. This study sought to identify individual differences that meaningfully predict one’s ability to detect these forgeries. It was hypothesized that measures of affect detection (theory of mind ability) and political  orientation would correlate with performance on a deepfake detection task. Within a sample (N = 173) of  college undergraduates and participants from Amazon’s Mechanical Turk platform, affect detection ability was shown to correlate with deepfake detection ability, r(171) = .73, p < .001, and general orientation to the political left was shown to correlate with deepfake detection ability, r(171) = .31, p < .001. Stronger correlations with deepfake detection ability were observed among specific facets of political orientation: economic liberalism, r(171) = .4, p < .001, and social progressivism, r(171) = .57, p < .001. However, affect detection ability was shown to mediate the relationship between deepfake detection ability and political orientation (Sobel Statistic = 5.29, SE = 3.29, p < .001). The deepfake detection task was also assessed as a predictor of an autism spectrum disorder screening instrument, r(171) = -.23, p < .001. The results of this study serve to identify populations who are particularly susceptible to deception via deepfake video and to inform the development of interventions that may help defend the vulnerable from nefarious attempts to influence them.