Jane Prophet

Artificial intelligence is an entangled, more-than-human relational network, shaping and being shaped by the societal, cultural, and political structures in which it is embedded. This paper explores the role of artists as critical practitioners engaging with AI, to examine how AI-generated self-representations materialise identity and reinforce or counter well-known AI biases in gender, race, and embodiment. Drawing on feminist technoscience – particularly its focus on the entanglement of body, environment, and technology – and autotheory, the study treats generative AI as an instrument of vision and voice. Using generative AI to create a partial digital twin, a self-portrait, is situated as an inherently embodied and relational practice. Through a combination of desk-based research and a practice-based, reflexive engagement with RunwayML, the paper documents the author’s attempt to create an identical-looking digital twin, revealing systemic biases embedded in AI-generated self-portraits. The paper uses embodiment as connective tissue linking theoretical and lived experience. Generative AI consistently misrepresented gender and age, defaulting to hyper-feminized aesthetics and youthful features while reliably reproducing Whiteness. The study also critically examines voice cloning and text-to-speech synthesis, highlighting how AI’s training data constrain language, accent, and vocal traits. By positioning AI-generated imagery and voice synthesis as material-discursive practices, the research extends debates on bias, agency, and self-representation in human–machine interactions. It argues that artists working with generative AI not only expose its epistemic limitations but also provide counter-narratives through creative, embodied interventions. The findings highlight the ways artists can help to meet the urgent need for more inclusive AI infrastructures, transparent dataset practices, and a reframing of digital self-representation beyond generative AI’s algorithmic defaults.