The Promise and Pitfalls of LLMs in Poetry Translation: A Case Study of Li Bai's "invitation to Wine"

The rise of Generative Artificial Intelligence is reshaping the practice of poetry translation. Using Li Bai's "Invitation to Wine" as a case study, this research compares an English translation generated by ChatGPT with the classic version by Xu Yuanchong. Through close reading across the dimensions of structure, culture-loaded terms, and rhythm, the study finds that while LLMs can produce grammatically correct and semantically coherent translations proficiently, significant gaps remain compared to human artistry. ChatGPT's output tends to be prosaic in form, literal in handling cultural imagery, and deficient in conscious rhythmic design.