Yin, HaoHaoYinDr. LIU Jianwen, KaceyKaceyDr. LIU Jianwen2025-08-062025-08-062024Yin, H., & Liu, J. (18 Apr 2024). Exploring gendered body language in AI translation: A corpus-assisted study. 2nd International Conference Translation and Cultural Sustainability, University of Salamanca, Spain.https://traduccioneinterpretacion.org/wp-content/uploads/2024/04/II-Congreso-TSC-Abstracts.pdfhttp://hdl.handle.net/20.500.11861/24396Gender bias in translation perpetuates stereotypical gender impressions, disadvantaging individuals of various genders. The rise of large language models and generative AI, notably ChatGPT, has increased the reliance on AI for supposed neutral and efficient translation. <br> This paper examines how ChatGPT’s translation reflects gender perspectives through gendered body language and investigates the potential for AI systems to echo and exacerbate existing human biases, as observed in popular machine translation tools like Google Translate. Using a corpus linguistics approach, this study compares the textual patterns of gendered body language in ChatGPT’s translation of “Chenzhong De Chibang” (1981) by Chinese female writer Zhang Jie, with translations by female translator Gladys Yang, “Leaden Wings” (1987), and male translator Howard Goldblatt, “Heavy Wings” (1989). Our findings indicate that ChatGPT, female, and male translators construct gender identities differently in terms of gendered body language. The analysis suggests that ChatGPT’s translations may reinforce traditional, non-inclusive gender perspectives, and underscores the need to integrate gender awareness into AI technologies and critical assess of AI translation.enConference Paper