기계번역의 한계와 번역사의 전략 - 유엔 총회 한·일 정상 연설문을 중심으로 -

Abstract

The purpose of this study is to examine the current level of machine translation by comparing it with the accuracy of human translators and hence to figure out the translation strategies that only human translators is are able to pursue. Specifically, error types of machine translation are analyzed in order to see the quality of artificial neural network-based machine translation, which is represented by Google Translation and Naver Papago. The analysis shows although the accuracy of machine translation has been considerably improved, it still generates many more errors than human translators. While comparing the two machine translations, Google Translation had more errors than Naver Papago Translation. When examining the form and strategy of human translators’ involvement in translation, machine translation still failed to follow the accuracy and precision of the translation of the strategic human translators. However, when it comes to the speed factor, machine translation was unparalleled. It was excellent. But no matter how advanced machine translation is, it most certainly has some limitations. For instance, it will fail to understand the culture, social trends, and current social phenomena, which reflect the tastes of readers. As a result, the mutual complementary efforts between machine translation and human translation will consistently be needed.