<![CDATA[This study investigates lexico-syntactic errors in Indonesian translations of English-language ELT journal abstracts, comparing outputs produced by human translators and AI-powered systems. The corpus consists of ten English abstracts and their corresponding translations, resulting in twenty texts analyzed in total. Human translations were prepared by graduate students trained in translation studies, while AI translations were generated automatically using neural machine translation tools. Each abstract was examined line by line, and translation segments were coded for lexical choice errors, syntactic errors, omissions, and additions. The analysis revealed sixteen dominant error instances, with human translations exhibiting a higher frequency of lexical and syntactic deviations, whereas AI translations were generally smoother but occasionally inaccurate in terminology and clause-level alignment. No omissions or additions were observed across the corpus, indicating that deviations were confined to word selection and sentence construction rather than content completeness. These findings highlight the importance of precise lexical choices and syntactic structuring in academic translation and emphasize the need for human post-editing when utilizing AI-assisted translation in the context of English Language Teaching. ]]>
Copyrights © 2026
