Exploring Cultural Sensitivity in Translation through a Comparison of AI and Human Capabilities

Abstract

Artificial Intelligence (AI) has become increasingly influential in translation, mainly because it can deliver translation services quickly and with relatively little effort from users. DeepL is one of the AI-based systems that has attracted attention for its ability to produce fluent and grammatically acceptable output. Even so, questions remain about how well such systems preserve cultural meanings, especially in literary works that contain local values, symbolic language, and culturally specific references. This study therefore compares DeepL-generated translation with human translation in the translation of Acehnese literary works, namely Hikayat, Hadih Maja, and Syair. Using a descriptive qualitative comparative design, the study examines 36 selected cultural expressions related to Islamic values, traditional titles, metaphors, symbolism, and local wisdom. Venuti’s (1995) concepts of foreignization and domestication provide the main analytical framework, while descriptive frequency counts are used to support the qualitative interpretation. The findings show that both DeepL and the human translator use foreignization and domestication, although clear differences can be seen in the way cultural meaning is preserved. The human translator more consistently retains Islamic terminology, cultural identity, metaphorical language, and symbolic expressions through context-sensitive interpretation, whereas DeepL more often renders cultural expressions in familiar English forms. These choices sometimes reduce religious, historical, and cultural meanings in the source texts. The differences are especially visible in Syair, where metaphorical language and implicit cultural references require interpretation beyond direct lexical correspondence. Overall, the study argues that cultural sensitivity in literary translation depends on more than linguistic accuracy and requires close attention to the cultural context of the source text. AI-based translation can assist translation work, but human translators remain important when culturally dense literary material must be preserved.