AI In Poetry Translation: Can Machines Capture Poetic Essence?

Poetry translation is a challenging process that extends beyond literal meaning, requiring sensitivity to emotional depth, cultural context, and poetic structure (Venuti, 2017). While artificial intelligence (AI) has advanced in machine translation, its ability to preserve poetic essence remains debated (Toral & Way, 2018). The purpose of this study is to analyze AI’s role in poetry translation, assessing whether it can retain the artistic and emotional depth of poetry or merely provide word-for-word translations. AI’s strengths lie in efficiency and linguistic accuracy, yet its limitations in handling metaphor, rhythm, and cultural context necessitate human intervention (Malmkjær, 2020). The significance of this study is twofold. First, it contributes to computational linguistics by evaluating how AI interacts with poetic language, a highly expressive and symbolic form of communication. Second, it provides insight into translation studies by assessing the extent to which AI can replicate human creativity in poetry translation (Tymoczko, 2007). The findings are relevant to translators, literary scholars, and AI developers aiming to enhance machine learning models for literary texts. By identifying the strengths and weaknesses of AI in poetry translation, this study advocates for AI-human collaboration, where human expertise refines AI-generated outputs for poetic authenticity (Garcia, 2021). Methodologically, the study employs a comparative analysis of AI-generated and human-translated poetry. Selected poems from multiple languages, including Nepali and English, will be translated using AI models such as Google Translate and DeepL. These translations will be evaluated based on linguistic accuracy, poetic aesthetics, and cultural sensitivity by literary experts (Toral & Way, 2018). Through qualitative assessment, the study aims to determine whether AI can go beyond literal meaning to capture poetry’s emotional and artistic essence.