Synthetic Identities, Real Consequences: Deepfakes, Gendered Representation and Digital Violence

DOI- https://doi.org/10.5281/zenodo.24197059

Abstract

The expansion of generative artificial intelligence has altered an old cultural problem: who has the authority to represent another person's body? Deepfake technologies make this question especially urgent because they can produce convincing images, videos and audio in which a real person's likeness is attached to events that never occurred. When such technologies are used to construct sexualised images of women without consent, the result is not simply misinformation. It is a form of gendered representational violence. This paper examines sexually abusive deepfakes through feminist approaches to visuality, surveillance, embodiment and narrative. It argues that the deepfake should be understood as an extension of image-based sexual abuse, but also as a distinct form of violence because fabrication itself becomes the instrument of harm. The paper brings Judith Butler's account of gendered performativity, Michel Foucault's discussion of surveillance, Laura Mulvey's theory of the gaze and Danielle Keats Citron's concept of sexual privacy into conversation with contemporary scholarship on technology-facilitated gender violence. It also considers the Indian legal context and the difficulties created by the speed and reproducibility of synthetic media. Rather than treating the problem as a narrow question of whether an image is authentic, the paper proposes a framework centred on consent, control over digital identity, platform responsibility, forensic capacity and narrative agency. It concludes that the humanitarian response to deepfake abuse must recognise the continuity between online representation and lived social consequences.

Keywords- Deepfakes, gender-based violence, artificial intelligence, image-based abuse, feminist visual culture, sexual privacy, digital identity, technology-facilitated violence, narrative agency, generative AI