Verifikasi Output Chatgpt Dalam Pembelajaran Bahasa Arab: Kesalahan Linguistik dan Validasi Kontekstual

The use of ChatGPT in Arabic language learning facilitates translation, text composition, grammar comprehension, and various linguistic tasks. However, the accuracy of its outputs is not always consistent, requiring verification before they are used as learning resources. This study aims to analyze linguistic errors in ChatGPT outputs, identify the need for contextual validation, and examine the involvement of humans and authoritative sources in the verification process. This study employed a Systematic Literature Review (SLR) using the PRISMA approach. Publications from 2022 to 2026 were retrieved through Google Scholar using Publish or Perish. Of the 100 publications initially identified, 10 studies met the criteria for analysis. The synthesis revealed grammatical errors, including the application of naḥwu rules and i‘rāb analysis, as well as morphological (ṣarf), semantic, and lexical errors, particularly in complex structures, specialized terminology, and translation tasks. Contextual validation is necessary to ensure that meanings align with textual, cultural, religious, and stylistic contexts. Based on these findings, this study proposes a four-stage conceptual verification framework comprising task identification, linguistic verification, contextual validation, and confirmation through human involvement and comparison with authoritative sources