Behind the Screens: Gendered and Generational Divides in Understanding Deepfake Violence

In 2024, all eyes turned to South Korea as it became the first victim of an epidemic of AI-generated deepfakes that disproportionately targeted women and minors. This technological threat, now surfacing worldwide, involves the non-consensual creation and distribution of hyper-realistic imagery, with current approaches to regulating this technology highlighting a societal unpreparedness and disconnect in understanding its true impact. This thesis thus explores the intricate issue of gendered digital violence, exploring how different groups perceive these harms, and the unexpected dual role of minors as both victims and perpetrators within this. Grounded in a virtual feminist theoretical framework, this multi-method qualitative research utilises focus groups and critical discourse analysis, with findings uncovering a 'Digital Violation Discrepancy' suggesting that the lack of women's perspectives, in AI development and regulation shapes understandings of deepfake harms. This discrepancy stems from a compounded issue: AI tools, instilled with patriarchal biases, birth an exploitative harm rooted in consent violation, further exacerbated by the crime's sui generis anonymity affordance that disrupts traditional legal and judicial proceedings relying on traceable evidence. These complexities were seen to be less understood by the male perspective, mirroring a critical gap in regulation and development measures that reflects this underrepresentation of women in these spheres. Consequently, deepfake creation's primary consent violation remains inadequately addressed in regulation, reflecting news representation where platform accountability is lacking, and fabricated harms, such as AI-generated child sexual abuse material, are normalised due to a misunderstanding of digital native behaviours. To address this, the research advocates for a foundational paradigm shift that prioritises women's experiences across all phases of AI development, ethical deliberation, and regulatory oversight, spanning AI governance and development reform, strengthening societal and educational responses, and encouraging international cooperation to harmonise legal frameworks. Ultimately, this thesis stresses that achieving an authentically equitable and secure digital future, safe from the uniquely gendered harms of deepfakes, requires challenging existing power structures within technology, ensuring AI empowers rather than exploits, particularly for women and minors.