Analysis of the Types of Translation Errors by the Google Translator

Abstract

Artificial intelligence translation programs have achieved a high level of reliability with respect to the accuracy of their translation output. In this regard, the goal of this research is to investigate to what level the AI translation program, especially the Google translation program, has advanced so far, and how soon the translation program will take us to the next level. To test the accuracy and reliability of the Google translator’s output, numberous Korean data were used for input including those which were carefully made up or collected to elicit lexical and syntactic ambiguities. The tests conducted in this study are aimed at identifying the types of errors that the Google translation program produces. The findings of the tests show that the Google translator makes various types of errors including lexical, syntactical and idiomatic ones. Lexical and syntactical errors are found to be made in the translation of ambiguous sentences in which contexts human translators are unlikely to make any translations errors, which is attributable to the Google translator's inability to understand non-linguistic cues or information. These results indicate there still remain a number of areas in translation that require human intuition and intervention.