Abstract
Artificially-Generated Non-Consensual Intimate Imagery (AI-NCII) is a form of gender-based violence and sexual abuse. Modern AI tools enable the easy creation of highly realistic NCII, with the potential to cause very real and profound harms to those depicted. Although victim characteristics, such as gender and sexual activity, are known to influence third-party judgements, less is understood about how victims from stigmatised or marginalised groups are judged and how these characteristics influence victim blame and perceptions of harm. In this preregistered vignette study, participants were presented with a hypothetical scenario describing a transgender or cisgender sex worker or office worker who was the victim of AI-NCII creation and subsequent sharing. The ideal victim framework and expectancy violations theory were used to interpret who was blamed and perceived as harmed, and why. Results revealed that sex workers were blamed more and perceived as less harmed than office workers. It was also observed that while transgender victims were blamed more by men than women, they were perceived as experiencing comparable levels of harm to cisgender victims. The findings suggest when victims do not conform to expected norms or are perceived to engage in behaviours inconsistent with ideal victim expectations, public sympathy may be reduced. Future interventions should therefore prioritise education about victimisation and challenge the expectations and beliefs that shape judgements of victims, rather than place responsibility on victims to conform to an ideal or reduce the visibility of their identities.