Automated Scoring of Arabic Text Using Large Language Models: A Literature Review

The article presents a literature review on Automated Text Scoring (ATS) for Arabic texts using Large Language Models (LLMs), examining both short answer grading (ASAG) and essay scoring (AES). It introduces a structured taxonomy that categorizes existing methodologies based on application domain, feedback generation, LLM architecture, competency alignment, and prompt engineering, while analyzing their datasets and performance metrics. This work underscores the importance of advancing ATS in Arabic to enhance educational assessment and quality in Arabic-speaking regions, emphasizing the need for further pedagogically oriented research in this field.