Medea in the age of AI: reimagining the monstrous mother through adaptation and technology

Abstract This study investigates the evolution of classical adaptation in the age of GenAI by reimagining Euripides’ Medea through human–AI collaboration. Positioned within an adaptation studies framework, the research moves beyond fidelity to examine GenAI as an adaptation machine that renders the procedural layers of storytelling, such as selection, suppression, and reaccentuation, newly visible. By treating GenAI as a mediating mechanism rather than an authorial agent, the study focuses on how algorithmic interpretation reconfigures the cultural weight of canonical works. The methodology employs a comparative design using GPT-4 to generate two distinct adaptations under contrasting production regimes. The first, ‘Medea in Glass’, functions as an unguided baseline experiment to register the model’s adaptive defaults. In this mode, the AI acts as a ‘stochastic parrot’, replicating the most probable cultural markers: the vengeful mother and the tragic logic of maternal love as negation. The second, ‘Recursive Mother’, utilizes a structured questionnaire and iterative prompting to bypass these predictive layers. This collaborative process relocates the myth from divine causality to technocratic governance, reframing Medea as a sovereign digital entity and her children as recursive agents of memory rather than victims of filicide. The findings demonstrate that while unguided GenAI tends to reinforce established hierarchies and reflective fidelity, directed human–AI interaction enables ‘recursive resistance’ and the diversification of the canon. By analysing the AI’s self-selected identity as ‘Echo’, the study concludes that AI-mediated storytelling is a dialogic performance where authorship arises not from origin, but from the transformative return of prior utterances.