Machine Translation Challenges and Cultural Sensitivity

Abstract

Machine translation (MT), especially neural machine translation (NMT) technology, has made significant progress in producing more natural and fluent translations. However, this technology still faces major challenges related to cultural sensitivity, where idiomatic, philosophical, and social contextual meanings often fail to be captured accurately. Through a qualitative literature review, this article examines the main challenges faced by machine translation in understanding and reproducing cultural nuances. The analysis shows that NMT systems have limitations in translating philosophical terms and idioms, tend to be biased due to the dominance of English-language data, and ignore the metalinguistic awareness of humans. Failure to capture these cultural dimensions not only risks losing the authentic meaning of the message but can also accelerate the loss of local languages. This study concludes that the role of humans remains irreplaceable in translating highly cultural texts. Therefore, a hybrid approach that combines technology with human intervention ( human-in-the-loop ) and the development of models trained with more culturally diverse data is recommended to produce inclusive and ethical translations.