Judgements of Deepfake Sexual Abuse Victims Differ as a Function of Facial Versus Body Likenesses

We are witnessing exponential growth in the use of machine learning to create fake – yet indistinguishable - sexual media of others without their consent. Though there is an emerging understanding of the impact that deepfake sexual abuse (DSA) has on its victims and societal understanding thereof, this knowledge pertains entirely to individuals whose facial likeness has been used within said material, with little-to-no attention paid to those whose bodies are used as the canvas; individuals who are predominantly sex workers. Across 321 participants (Mage = 45.70 years, SD = 15.88; 48.9% female), vignettes were used to explore differences in societal judgements of DSA for victims whose face (versus body) was used to generate DSA material, and whether they were labelled as a sex worker. Though perceived criminality did not differ across conditions, participants allocated more blame and less anticipated harm to DSA victims whose body, relative to whose face, was used. This effect was enhanced in vignettes labelling them as sex workers. When exploring correlations using demographics, beliefs, and personality traits, being older, male, and viewing sex work as ‘a choice’ and/or ‘deviant’ predicted greater victim blame, lower perceived criminality, and less anticipated harm. High self-reported empathy was the only predictor of greater anticipated harm. Results indicate the importance of understanding broader impacts of DSA - regardless of the victim - for stakeholders within the criminal justice system, and a continued need to generate public awareness of DSA through international policy.