COMPARISON OF INSTAGRAM TRANSLATE AND HUMAN TRANSLATION IN THE ENGLISH CAPTIONS OF JOKOWI’S INSTAGRAM ACCOUNT

Abstract

Translation technology has progressed so swiftly that it may soon replace human translation. When it comes to translation, artificial intelligence-based translation machines are increasingly resembling humans. Instagram Translate (IG Translate), for example, has shown a steady improvement in translating user-posted photo captions. Many studies have been conducted to investigate the performance of IG Translate in various language pairs. Instagram began offering a translation feature for its users in 2016 in order to translate the photo captions posted by users. There is a chance that the translations contain errors, which should be investigated in order to assess the translation performance of IG Translate. The purpose of this research is to look into Koponen's translation errors category as performed by IG Translate in translating the photo captions on Jokowi's official Instagram account. However, whether IG Translate outperforms a professional human translator remains to be seen. This study aims to assess the translation performance of IG Translate and human translators by identifying potential errors and analyzing the translation strategies used by the machine and the human. The translation error analysis is expected to determine whether the machine and the human use different or the same translation strategies. The photo captions used for the analysis were taken from Jokowi's official Instagram account between March and September of 2020. This range was chosen because it represents the start of the Covid-19 pandemic as well as the government's approach to overcoming the pandemic. The research is expected to provide a theoretical benefit by expanding the translation research repertoire, as well as a practical benefit for the IG Translate developer to improve its algorithm and for social media users to input their photo captions in a way that the machine can translate well.