FROM WORDS TO ERRORS: ANALYZING LEXICAL CHALLENGES IN INSTAGRAM'S MACHINE TRANSLATION OF 'NYTIMES’

In the current digital era, disseminating information via social media is very easy and quick, but this can also increase the spread of wrong and inaccurate information to the public caused by translation errors in automatic translations on social media. The main objectives of this research are 1. Identify the types of translation errors found in Instagram Machine Translation in Translating the Caption of the "NYTimes" Instagram Account. 2. search for the most dominant types of lexical errors made by Instagram Machine Translation in Translating the Caption of "NYTimes" Instagram Account. The theory of error classification by Vilar et al. (2006) is applied to identify three main types of errors: unknown words, missing words, and incorrect words. The research method involves analyzing 100 translated news captions. There are three types of Lexical Errors found in Instagram machine translation when translating captions from the NYTimes Instagram account (1) Incorrect Word (4 instances, 50%) (2) Omitted or Missing Word (2 instances, 25% (3) Unknown Word (2 instances, 25%). The most common type of error identified in the translations is the use of incorrect words, accounting for half of all errors. This indicates that the machine translation often fails to select the appropriate words to convey the original meaning effectively.