Accuracy of DeepL Translation in Inside Out 2 Movie Compared to Official Subtitles

Abstract

This study aims to describe the level of translation accuracy produced by DeepL in the film Inside Out 2 and compared it with the official subtitles. Using a qualitative descriptive method, the researcher selected 30 segments from the film's dialogues that contain various expressions, idioms, and contextual meanings. The accuracy of DeepL translations was assessed using the translation quality assessment model developed by Nababan, Nuraeni, and Sumardiono (2012), which classifies translations into three levels: accurate, moderately accurate, and inaccurate. The results showed that most of DeepL's translations felt into the "moderately accurate" category with total 67%, 17% as accurate and 17% also as inaccurate level. This was due to literal translation, misinterpretation of idiomatic expressions, and lack of contextual adaptation. In contrast, the official subtitles demonstrated higher accuracy by delivering contextually appropriate and culturally sensitive translations. These findings suggested that while DeepL offers grammatically fluent output, it still lacks the semantic and pragmatic sensitivity required for high-quality subtitle translation in films.