Prompting Poetry: Assessing ChatGPT 4o’s Capacity in Poetry Translation

Abstract

This study investigates the potential of generative artificial intelligence in the translation of poetry, focusing on Sylvia Plath’s “Child” as a central case. Guided by André Lefevere’s strategies, seven distinct prompts were designed to generate English-to-Turkish translations, aiming to assess how AI responds to targeted instructions and whether it can engage with the specific demands of poetic language. The research employs a qualitative methodology to evaluate the extent to which AI can address the challenges of literary translation. With this approach, the study provides a perspective on how poetry, traditionally viewed as one of the most translation-resistant forms, is reinterpreted within a digital environment. The analysis assesses AI’s performance in terms of prompt responsiveness. Accordingly, it highlights the relationship between artificial intelligence and strategic translation processes, contributing to ongoing discussions at the intersection of technology, translation, and literature.