Hanifah Miladina Isnaini; R. Nabila Zaty Shakila
With the advancement of technology, artificial intelligence tools have begun to emerge in the field of education, such as ChatGPT, which is now used by some teachers at Pondok Modern Darussalam Gontor to compare translations of Indonesian texts into Arabic. Although there are recognized reference books for correcting translations, there are discrepancies between machine translations, including errors in terminology selection and contradictions with the educational context. This raises questions about the accuracy of ChatGPT’s translations, making this study necessary to analyze and evaluate terminological errors. This study aims to analyze the translation errors made by ChatGPT when translating Indonesian articles into Arabic. This study also seeks to evaluate the quality of these translations based on the classification of terminological errors and Ali Al-Qasimi’s theory regarding the evaluation of terminology in terms of acceptance and rejection. This study employs a qualitative approach using content analysis, in which ChatGPT’s translations are compared with those in reference books, with reference to specialized Arabic dictionaries. Subsequently, the data was analyzed to identify error patterns and their acceptance levels within an educational context. The results indicate that translation errors can be classified into four main patterns: phonetic transfer, conceptual errors, semantic shifts, and inaccurate synonyms.