Engaging With <scp>AI</scp> Tools in Translation Tasks: Chinese University <scp>EFL</scp> Students' Perceptions and Practices

Engineering notes

Engineering notes will be added by the aipentium editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为大语言模型、生成式AI、ChatGPT相关技术、计算机视觉、深度学习等高价值论文补充中文说明。

Original abstract

ABSTRACT Background The rapid advancement of artificial intelligence (AI) is reshaping language education, with AI‐assisted language learning (AILL) offering new opportunities for personalised support and interactive engagement. However, it remains underexplored how English as a Foreign Language (EFL) learners perceive and engage with AI tools in specific learning contexts, particularly in translation tasks. Objectives This study explores Chinese EFL learners' perceptions and practices in AI‐assisted translation tasks, with a particular focus on the interplay between perceptions and practices, as well as the factors shaping this engagement. Methods The study was conducted at a provincial “Double First‐Class” university in China and involved third‐year English majors enroled in the compulsory course Advanced English , where students completed authentic Chinese‐English and English‐Chinese translation tasks using AI tools, including ChatGPT, DeepL and Wenxin Yiyan. Data were collcted from six participants through learning journals completed over one academic semester and post‐course semi‐structured interviews. Results and Conclusion The findings suggest that participants in this study tended to express generally positive orientations toward AI tools, particularly in terms of enhancing translation efficiency and providing access to diverse linguistic resources, while also identifying limitations related to cultural contextualisation, fluency and the accuracy of specialised terminology. Learners appeared to follow an emergent three‐stage pattern of engagement consisting of draft generation, multi‐tool cross‐validation and manual optimisation, reflecting a critical and strategic rather than uncritical use of AI. This process was mediated by factors including language proficiency, task complexity and the technological features of the employed AI tools. The study contributes to current understandings of AILL by offering qualitative insights into how learner agency is reconfigured in AI‐mediated translation tasks and by highlighting translation as a productive context for examining both the opportunities and constraints of AI integration. The findings also provide pedagogical implications for supporting more critical, reflective and context‐sensitive uses of AI in language learning environments.

Links and sources

Comments