Analyzing the Performance of Gemini, ChatGPT, and Google TranslateinRendering English Idioms into Arabic

This study examines the translation of 155 idioms by different machine and AI translationsystems, namely Google Translate, ChatGPT, and Gemini. Various sources were utilizedtocollect the data, including books, magazines, interviews with native English speakers, andvarious websites dedicated to English idioms. This data was analyzed based on a frameworkbuilt on the taxonomy of Baker (1992). The quantitative part examined the frequencyoftranslation approaches each program used to render the idioms. The qualitative part focusedonselected examples to highlight the potential issues of each approach in conveying the styleandsense-based features of the idioms. The findings showed that idiom translations weredonethrough three main approaches: literal, sense-based, and idiom-to-idiomtranslation. GoogleTranslate had the highest percentage of literal translation at 76%, followed by ChatGPTat 53%, while Gemini had the lowest percentage at 21%. For sense-based translations that usenonfigurative language, Gemini was in the lead at 63%, followed by ChatGPT, with a widegapat 35%. Google Translate had the least sense-based renditions at a mere 11%. When it cametotranslating idioms using figurative language, Gemini once again was in the lead with16%, followed closely by ChatGPT at 13%, with Google Translate right behind at 12%. The studyconcludes that although there is vast improvement and advancement in technology, machinetranslation has yet to master nonliteral language such as idioms.