A Scoping Review of Artificial Intelligence-Generated Image-Based Sexual Abuse: Terminology, Nature, Prevalence,...

Artificial intelligence-generated image-based sexual abuse (AI-IBSA), colloquially known as "sexual deepfakes," includes the non-consensual creation and/or sharing of intimate images generated by and/or modified using artificial intelligence ("AI-IBSA content"). While literature on AI-IBSA is growing, a scoping review of empirical (scholarly and gray) literature on the terminology, nature, prevalence (victimization and perpetration), community attitudes toward, and harms and impacts of AI-IBSA by and/or against adults is yet to be undertaken through a gender-inclusive lens. Thus, we searched six bibliographic databases to identify studies that met seven inclusion criteria: (a) empirical research; (b) on AI-IBSA perpetration and/or victimization; (c) of adults; (d) using qualitative and/or quantitative methods; (e) published as scholarly or gray literature; (f) between January 2017 and January 2026 (inclusive); and (g) in English. Across the 34 reviewed studies, our critical findings include that most terminology lacked consistency and clarity, did not explicitly name AI-IBSA as abuse, and conflated AI-IBSA content with pornography in general. In terms of gender, content analyses showed that publicly available AI-IBSA content on selected websites overwhelmingly appears to depict women, but prevalence studies painted a more complex picture, with men self-reporting higher rates of victimization for some AI-IBSA behaviors than women. Despite the significant harms of AI-IBSA, community attitudes differed based on victim-survivor type, perpetration motivations, and/or other contextual factors. We argue that AI-IBSA research, policy, and practice could particularly benefit from further prevalence studies, diverse lived experience expertise, and prevention responses that aim to dismantle the harmful norms and attitudes undergirding this form of technology-facilitated gender-based violence.