Abstract
In the context of globalization, the translation quality of publicity texts on world heritage sites influences China's cultural dissemination. However, systematic comparisons between human and AI-generated translation in this field remain scarce. To this end, this study employs quantitative linguistic methods and builds a Chinese-English parallel corpus based on China's world heritage sites. It compares the official human translation with a version produced by DeepSeek. Using Wordsmith, TreeTagger, and Paraconc, the analysis looks at lexical richness, term consistency, high-frequency word patterns, average sentence length, and syntactic features, combining statistical data with qualitative textual analysis. The results show that AI translations tend to appear more varied at the lexical level, but they are less stable in key terminology and sometimes inconsistent in meaning. At the discourse level, they favor shorter sentences and more explicit connectives, which improves readability but reduces information density and weakens textual coherence. These problems stem from AI's limited domain modeling. The study advocates for an enhanced terminological control and human-AI collaboration to achieve both efficiency and quality in external publicity translation.
References
Baker, M. (1995). Corpora in translation studies: An overview and some suggestions for future research. Target, 7(2), 223-245. https://doi.org/10.1075/target.7.2.03bak
Baker, M. (2000). Towards a methodology for investigating the style of a literary translator. Target, 12(2), 241-266. https://doi.org/10.1075/target.12.2.04bak
Feng, Q. H. (2012). Translations and vocabulary under the thinking mode. Shanghai Foreign Language Education Press.
Huang, Y. Y. (2004). Adhering to the “Three Closeness” principles in international communication and addressing difficult issues in translation for international audiences. Chinese Translators Journal, (6), 29-30.
Lu, X. J. (2013). National image and translation strategies for international communication [Doctoral dissertation, Shanghai International Studies University].
Lu, Y. Q. (2026). A comparative analysis of DeepSeek translation and human translation: A case study of The Pole. Jiaying Literature, (1), 114-116.
Wang, H. S., & Liu, S. J. (2022). A study on post-editing competence of machine translation in the era of artificial intelligence. China Science & Technology Translators Journal, 35(4), 21-24. https://doi.org/10.16024/j.cnki.issn1002-0489.2022.04.010
Wei, Y. M., Jia, K., Zeng, R. X., He, Z., Qiu, L., Yu, W. X., ... & Jiang, Y. H. (2025). Artificial intelligence innovation development and governance transformation under the DeepSeek breakthrough effect. E-Government, (3), 2-39. https://doi.org/10.16582/j.cnki.dzzw.2025.03.001
How to cite this paper
Human vs. AI Translation of Chinese Publicity Texts: A Corpus-based Comparative Study Drawing on China’s World Heritages
How to cite this paper: Jinghan Xie, Jilin Fu. (2025). Human vs. AI Translation of Chinese Publicity Texts: A Corpus-based Comparative Study Drawing on China’s World Heritages. Translation and Foreign Language Learning, 1(5), 801-805.
DOI: http://dx.doi.org/10.26855/tfll.2025.12.009