DH 2026

Daejeon, July 27–31

Thu, July 3016:30–18:00S078101-102
Short Paper

Demythologising through Stable Diffusion: AI-Generated Visualisations in Angela Carter’s The Passion of New Eve

Longlong Ge
University of Nottingham, United Kingdom · longlong.ge@nottingham.ac.uk
Lei Yan
University of Cambridge, United Kingdom; Cambridge Digital Humanities, United Kingdom · imruby070809@gamil.com

This project draws on Angela Carter’s notion of the ‘demythologising business’ ([1997] 2013: 27): a literary practice that dismantles dominant symbolic systems surrounding gender, sexuality, and identity—to explore how Stable Diffusion (SD), an open-source latent diffusion model, might be repurposed in the service of feminist literary criticism (Mehrabi et al., 2021). By reworking and fabricating alternative creation myths and odyssey narratives, The Passion of New Eve (1977) exposes, interrogates, and ultimately destabilises the patriarchal myths that underpin gender, identity, and power. Through demythologisation, Carter not only defamilarises the cultural fabrication of femininity, but also the processes through which identity is produced, performed and undone.

Building on this framework, our study employs AI-generated imagery to visualise female figures from The Passion of New Eve through a systematic process involving prompt design and refinement, iterative evaluation, parameter adjustment, the construction of an image lexicon, and manual editing. By juxtaposing the generated images with traditional gender iconographies and with Carter’s own narrative textures, we investigate how generative AI can extend literary demythologisation into the visual domain—transforming the act of seeing into an act of critique. Crucially, our inquiry does not end with the technical evaluation of image quality or stylistic coherence. We propose a feminist critical framework that integrates aesthetic assessment of images (fidelity, authenticity, and affectivity) with political interrogation on generating process (post-identitarian critique and discursive reframing). In doing so, we demonstrate that image generation can operate not merely as illustration but as a form of feminist critique. This project suggests a reciprocal relationship between feminist literature and generative AI: literature provides the ethical and conceptual grounding for critical AI practices, while AI enables feminist criticism to evolve into new multimodal, affectively charged terrains that exceed the limits of textual representation.

References
  1. Carter, Angela. (1977). The Passion of New Eve. London: Virago Press.
  2. Carter, Angela. ([1997] 2013). “Notes from the Front Line”. in: Shaking a Leg: Collected Journalism and Writings. New York: Penguin Books 26–30.
  3. Mehrabi, Ninareh, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan. (2021). “A Survey on Bias and Fairness in Machine Learning”, in: ACM Computing Surveys 54, 6: 1–35.