隨著 Open AI 開發之 ChatGPT 於 2022 年正式問世,人工智慧(Artificial Intelligence, AI)的技術發展再次來到新紀元,AI 的蓬勃發展大幅影響人 類社會,各行各業也因應開始做出調整。綜觀筆譯產業,「在地化翻譯 (Localization Translation)」可說是最頻繁使用翻譯科技工具的代表,本 研究將透過整理過去文獻以及產業資訊,詳細介紹本地化翻譯產業、電腦 輔助翻譯(Computer-Assisted Translation)以及 AI 相關翻譯技術,並透過 半結構式訪談,與本地化產業之專案經理(PM)、語言測試人員(LQA)、 審稿、資源管理專員、譯者等不同角色進行深度訪談,從業界不同崗位的 觀點,探討本地化翻譯產業的工作型態、翻譯科技使用現況、以及人類譯 者與機器翻譯之未來發展及可能性。 根據訪談內容有以下總結:(一)本地化之 QA report 並非定奪翻譯 水準的絕對標準,譯者應保持主動溝通;(二)與內部譯者相比,自由譯 者需維持跨時區工作的 work-life balance;(三)除了翻譯能力,合格譯者 最不可或缺的特質是「尊重交期」;(四)本地化翻譯與科技息息相關, 無論是 CAT 或是 AI,從業者都應抱持開放的學習態度。 最後,針對 CAT 與 AI 應用,本研究透過訪談發現:(一)過去研 究提及之 Trados 使用問題,現今已都解決;(二)AI 技術使得機器翻譯 愈加成熟,「譯後編輯」已成為不可擋之趨勢;(三)從業人員無需害怕 AI 取代人類,而是在人機協作的基礎上進行突破;(四)未來的 CATTool 可能會加強譯後編輯的功能,甚至透過 AI 訓練出譯者個人之機器翻譯。
In 2022, OpenAI's ChatGPT marked a new era in AI development, profoundly impacting society and prompting industry adjustments. This study focuses on "Localization Translation Industry," which frequently utilizes translation technology tools for translation projects. Through semi-structured interviews with key industry roles including 1 project manager, 2 linguistic quality assurance (reviewers), 1 resource management specialist, and 2 translators, the research comprehensively explores the industry's working patterns, current translation technology usage, and the future possibilities for human and machine translation. Based on the interviews, key conclusions are: (1) QA reports in localization are not the sole standard for translation quality, proactive communication is vital; (2) Freelance translators must balance work-life across time zones compared to in-house translators; (3) Meeting deadlines is crucial alongside translation skills; (4) Embrace CAT and AI with an open learning mindset for localization's tech- related nature. Regarding CAT and AI applications, the study finds: (1) Past Trados issues have been resolved; (2) AI has matured machine translation, leading to the growing "post-editing" trend; (3) Embrace human-machine collaboration for industry advancements; (4) Future CAT Tools may enhance their post-editing capabilities and even train individual machine translations through AI.