Redefining Grammatical Competence in the Generative AI Era: A Nahwu-AI Integrated Competence Model for Assessing and Ensuring Quality in Nahwu Teaching

Generative artificial intelligence and large language models have transformed the production and evaluation of Arabic grammatical texts, while assessment tools in Nahwu teaching still rely on competence models developed before this disruption. Existing models by Canale and Swain, Bachman, and the Common European Framework of Reference provide valuable foundations but do not specifically address Nahwu constructs such as case endings, 'amil-ma'mul relations, mu'rab-mabni distinctions, or AI-assisted i'rab analysis. This study employed library research to develop a Nahwu-specific competence model and quality assurance mechanism based on literature in applied linguistics, Arabic natural language processing, and educational assessment published between 2022 and 2025. The findings propose the Nahwu-AI Integrated Competence Model, which retains four classic communicative competencies while re-anchoring grammatical competence in core Nahwu constructs, including i'rab (rafa', nashab, khafd, jazm), 'amil-ma'mul relations, and mu'rab-mabni classification. A fifth dimension, Generative AI Interactive Competence, comprises prompt literacy, verification literacy, and reflective literacy, operationalized through concrete i'rab and 'amil-ma'mul tasks. The study also proposes three-level quality assurance, multi-level task design, dual-criteria calibration, and dialectal and stylistic bias review to support valid and fair Nahwu assessment.