Human expertise vs AI efficiency : a comparative analysis of student and ChatGPT post-editing

Abstract

This study investigates the differences between student post-editing (SPE) and ChatGPT-based post-editing (GPTPE) in Chinese-to-English translations of tourism texts. Utilizing a mixed-methods approach, the research combines quantitative analysis of linguistic features—including lexical diversity, lexical density, sentence length ratio, and noun-to-verb (NV) ratio—with qualitative insights from student reflection reports. Results reveal that GPTPE outputs exhibit higher lexical diversity and density, leveraging extensive linguistic resources to produce varied and information-rich translations. Conversely, SPE outputs demonstrate greater genre sensitivity, holistic perspective and cultural adaptation, with student translators employing strategies that prioritize the target audience’s expectations and the informative and persuasive functions of tourism texts. Despite lower lexical variety, students excel in contextualizing content and mediating cultural nuances. The findings highlight the complementary strengths of AI efficiency and human expertise, underscoring the need for translation education to foster critical evaluation skills, creativity, and user-centered approaches. This will better prepare future translators for effective collaboration with AI technologies in the evolving landscape of the translation industry.