Èkphrasis Reimagined: The Impact of AI on Interpretation and Generative Meaning

This study examines the transformation of èkphrasis in the digital era, specifically regarding artificial intelligence AI-powered generative systems. Èkphrasis, originally defined as a vivid verbal represen-tation of visual art, has transcended its classical rhetorical and literary roots to encompass transme-dia applications facilitated by computational technologies. The research examines how AI-mediated text-to-image generation alters ekphrastic practice, shifting from descriptive narration to a process of interpretative co-creation between human users and Machine Learning models. Generative AI, through probabilistic associations and multimodal synthesis, challenges conventional notions of artistic agency, requiring a reevaluation of the relationship between linguistic input and visual depiction. This paper analyzes the balance between descriptive and evocative language in AI prompting, highlighting the influence of algorithmic biases, latent-space dynamics, and iterative refinement in creating an in-teractive ekphrastic framework. Ultimately, it argues that AI-assisted ekphrasis transcends traditional boundaries of textual and visual interpretation, positioning itself as a creative, curatorial, and compu-tationally mediated process of meaning production, where the role of the human creator is centred in taking responsibility for the guidance towards, and selection of, the final outcome based on his/her creative vision.