Abstract
This thesis discusses the third presidential candidate debate on KPU RI official youtube channel. The aim of this research is to investigate the types of euphemisms, focused on identifying the translation techniques of euphemisms found, and showed the translation accuracy result in the automated subtitle translation by MT provided in the presidential debate video on the KPU RI YouTube channel. The research uses Allan and Burridge’s (2020) theory, the types of translation techniques by Molina and albir’s (2002) theory, and the translation accuracy by Nababan in Ardi (2016). To conduct this study, the researcher use descriptive-qualitative method. Based on the findings, It found that there are 3 types of euphemism in the source text, they are analogy, distortation, and borrowing. The highest frequency of euphemism found in the transcript of KPU RI Youtube “Debate of 2024 Presidential Candidates” was distoration with the frequency 42,57% from 148 data and the highest frequency of translation techniques was literal translation with frequency 57,33% from148 data.
Key words: Translation, Euphemism, Machine Translation