Linguistic and Stylistic Over-Elaboration in AI-Permitted Arabic Writing Tasks: Evidence from Beginner Learners

This study examines linguistic and stylistic over-elaboration in beginner Arabic writing produced during a classroom task permitting generative artificial intelligence (GenAI) and electronic translation. The study emerged from an Arabic language course involving undergraduate accounting students at Universiti Teknologi MARA (UiTM), Terengganu Branch, Malaysia. Students were instructed to write a simple birthday invitation addressed to friends, using accessible vocabulary and including specified invitation elements. During the review of approximately 100 submissions, the instructor observed recurring instances of vocabulary, courtesy expressions, and sentence formulations that appeared more sophisticated or formal than expected for the learners' proficiency and the task's communicative purpose. Adopting an exploratory qualitative approach, the study draws on these professional classroom observations, supplemented by the direct examination of three available invitation texts. The analysis considers lexical overreach, stylistic over-elaboration, unnecessary syntactic complexity, register mismatch, and learner-level appropriateness. The findings highlight a pedagogical distinction between linguistic correctness and the suitability of written expression for a beginner's level and communicative situation. They also suggest that apparently sophisticated writing should not automatically be interpreted as evidence of independent linguistic proficiency. Because individual tool use and writing processes were not documented, the study does not attribute particular expressions to GenAI or establish causal effects.