Enhancing Literary Translation With Generative AI: A Stylistic Approach

This study explores the integration of stylistics into a generative AI-assisted “Introduction to Literary Translation” course. Using the 5E instructional model, students are guided to optimize translation prompts for novels and poetry, evaluate machine translation (MT) quality, and conduct post-editing to enhance their mastery of stylistic features and post-editing skills. The research questions include: (1) Whether incorporating stylistics into prompts enhances students’ ability to preserve the original stylistic features, (2) The impact of GenAI translation on creativity, and (3) Students’ perceptions of AI-assisted literary translation. This study adopts an action research approach, utilizing pre- and post-tests, questionnaires, and interviews for analysis. Preliminary findings suggest that AI can assist in preserving stylistic features, particularly in pragmatics, tone, rhetoric, and rhyme schemes. However, some students believe AI constrains creativity, requiring multiple rounds of post-editing to retain the original style. Additionally, while most students acknowledge that AI improves efficiency and consistency, they also emphasize the need to avoid over-reliance on AI. This study recommends that translation courses strike a balance between AI assistance and human translation to ensure both quality and creativity. This research underscores the importance of integrating AI literacy into translation education. Students need to understand the capabilities and limitations of AI tools, learning how to leverage them effectively while maintaining their own linguistic and creative agency. A blended approach that combines AI support with traditional translation methods can foster both efficiency and depth in literary translation training.