Assessment of Cultural Sensitivity in the ChatGPT’s Translation of the Acehnese Chronicle ‘Ali Peureudan’

##plugins.themes.bootstrap3.article.main##

Sabrina Sabrina Ghazali Takary

Abstract

Cultural sensitivity is an integral aspect in translation to ensure the preservation of context and meanings, allowing the readers to grasp the intended meanings; however, this aspect was often omitted and still highly under-researched. The research aimed to examine cultural sensitivity in the ChatGPT’s English translation of the Acehnese Chronicle of Ali Peureudan originated from Aceh Province, Indonesia. The story contains culturally-specific phrases in Acehnese and Arabic languages. The research was a Comparative Literary Analysis which is descriptive qualitative in nature.  The source text (ST) and Target Text (TT) generated by ChatGPT were compared and analyzed carefully using the Translation Quality Assessment developed by Juliane House (2015). There were 26 cultural phrases in total under analysis. The results revealed that 85% of ChatGPT’s renditions of culturally-bound terms were covert, which means that the phrases were adapted to the target language norms (English), such as Allah as ‘God’, Insya Allah as ‘God willing’, and ‘Surat’ as ‘Wish’, causing a loss of cultural nuance, generalization, semantic narrowing, and meaning distortions. Only 15% of the analyzed phrases were translated overly by maintaining their original versions, avoiding diminution of cultural meanings. Of the 15%, only a single word was found to be kept intact in its original version yet accompanied by a meaning description, namely the word Alhamdulillah (Praise be to God).  Such strategy protects its originality, avoids erasure of cultural details, as well as provides understanding to the target readers of its intended meaning. This research concludes that ChatGPT still lacks cultural sensitivity in rendering a literary work. This research is expected to be a foundation for those using ChatGPT as a translation tool for such culturally sensitive texts, and provide an input to ChatGPT developers to improve its system for more accurate rendition.

##plugins.themes.bootstrap3.article.details##

Section
Articles