The Development and Validation of the Deepfake Myth Acceptance Scale (DMAS)

Since their inception, deepfake videos (i.e., AI generated or modified videos that depict a person saying or doing something they never did) have been primarily used to create non-consensual sexual imagery of women. We argue that false beliefs about the harms of deepfake technology (e.g., the idea that fake videos are not harmful because they are fake) encourage the watching, creation and sharing of deepfake content. We report on the development and validation of a 6-item deepfake myths acceptance scale (DMAS) over 3 studies (n = 1080) which consist of an Exploratory Factor Analysis, a Confirmatory Factor Analysis and a second Confirmatory Factor Analysis, which focused specifically on high-perpetrating populations. The scale has strong psychometric properties (Cronbach's α = .87) and moderately correlates with self-reported past behaviours and future intentions to watch, share, and create deepfakes. The scale takes approximately 2 minutes to complete and is a useful tool for researchers who seek to understand the individuals and communities that produce and share deepfakes. We also encourage researchers to utilise the deepfake myth acceptance measure to assess the effectiveness of interventions at changing the attitudes/beliefs that underpin deepfake abuse.