比較專業譯者與終端使用者對ChatGPT翻譯功能的態度

Abstract

2022年問世的生成式語言模型ChatGPT經常被大眾作為翻譯工具使用。而專業譯者與終端使用者使用機器翻譯系統的情境與目的可能不同，或許也會導致兩族群對ChatGPT翻譯功能的態度與接受度產生差異。本文旨在探討譯者與終端使用者使用ChatGPT翻譯功能的態度及使用意圖，研究結合Davis（1989）提出的第一版科技接受度模型及Venkatesh與Davis(2000)之第二版科技接受度模型為基礎，進行問卷調查及半結構式訪談。問卷結果顯示，大多數的譯者及終端使用者均使用過此模型的翻譯功能，而使用過此功能的使用者擁有較高的使用意圖，譯者多利用此模型產出之譯文以大致了解原文內容，終端使用者則是為了閱讀外語文本而使用。此外，使用者對ChatGPT翻譯功能的信任度會因不同語言組合及輔助功能而有所差異。兩族群的質性訪談顯示，終端使用者更傾向認為譯者會被AI取代，但此族群對翻譯產業的了解程度相對有限，多從效率與成本層面看待此議題。而譯者則更擔憂大眾對翻譯專業低估，可能對整個翻譯產業帶來惡性循環。儘管意識到生成式語言模型對翻譯產業的衝擊，多數譯者仍抱持相對樂觀的態度，認為譯者以開放心態擁抱新科技，方能於勢不可擋的AI浪潮中成功存活。ChatGPT, a generative language model released in 2022, is frequently used as a translation tool. However, professional translators and end-users employ machine translation systems in distinct contexts and for different purposes, which may lead to contrasting attitudes and acceptance levels towards ChatGPT's translation functions. This study aims to explore the attitudes and intentions of translators and end-users toward the translation function of ChatGPT. Grounded in Davis' (1989) TAM and Venkatesh and Davis' (2000) TAM2, the research employed a combination of questionnaires and semi-structured interviews. Survey findings revealed that the vast majority of both translators and end-users have utilized the translation function of this model. Moreover, participants with prior usage experience demonstrated higher usage intentions. Translators primarily utilized the model's output for general comprehension of source texts, while end-users employed it for reading foreign language documents. The study also found that user trust in ChatGPT's translation capabilities varied depending on language combinations and auxiliary functions. Qualitative interviews with both groups suggest that end-users were more inclined to believe translators will be replaced by AI. However, this group exhibited a relatively limited understanding of the industry, often approaching the issue from an efficiency and cost perspective. In contrast, translators are more concerned about the public's underestimation of the translation profession, which could potentially lead to a negative spiral for the entire industry. Despite acknowledging the impact of generative language models on the translation industry, most translators maintained a relatively optimistic outlook, emphasizing that embracing new technologies with an open mind is crucial for survival in the inevitable AI wave.