Gender Bias In Machine Translation (Google Translate) From Indonesian To English

Abstract

This study aimed to explain the gender translation pattern from Indonesian to English in machine translation (Google Translate) and the machine translation bias form in gender translation. The method used is descriptive qualitative with Sangkuriang folklore as the object. The primary data were Sangkuriang folklore translation result from Indonesian to English using Google Translate and the secondary data were books or references related to machine translation, language, and gender. Data were collected by documentation technique, document analysis, and note-taking technique. The results showed that Google Translate had a tendency to identify the word "dia" in Indonesian for the male gender "he" in English. Although the source language has provided context for the clause before the keywords 'he' and 'she', the machine translator still translated the word 'dia' in the source language into the word 'he' in the target language. Besides, Google Translate still embedded a positive gender depiction towards the male, namely by depicting brave, optimistic, tough, able to take risks, and likes challenges. Meanwhile, the depiction of negative traits was addressed to the female. In addition, it mapped the roles of men and women where men were as leaders because they are brave, tough, and like challenges; while women were as subordinate figures who were only able to choose, but did not have full power to determine and lead.