<![CDATA[Achieving grammatical accuracy in academic writing remains a persistent challenge for English as a Foreign Language (EFL) students, often constrained by the logistical limitations and delayed nature of traditional instructor-led feedback. While Artificial Intelligence (AI) offers a potential solution, empirical evidence comparing its instructional efficiency against human correction remains inconclusive. This study aims to investigate the comparative efficacy of AI-based Automated Corrective Feedback (ACF) versus traditional Manual Corrective Feedback (MCF) in improving grammatical accuracy. Employing a quasi-experimental design with a non-equivalent pretest-posttest control group, the research involved 60 undergraduate students at Universitas Kiai Abdullah Faqih, Indonesia. Participants were assigned to an Experimental Group utilizing Grammarly (N=30) and a Control Group receiving coded manual feedback (N=30) over a 12-week intervention. Grammatical accuracy was measured using the Error-Free T-unit ratio. The Analysis of Covariance (ANCOVA) results revealed that while both modalities yielded improvements, the AI-assisted group (M=84.20) significantly outperformed the manual feedback group (M=73.50) with a large effect size (p < .001, partial eta squared = .425). These findings suggest that the immediacy and non-judgmental nature of AI feedback accelerate the mastery of surface-level grammar. The study advocates for a 'Hybrid Intelligence' model, where educators leverage AI to handle mechanical corrections, thereby allowing human instruction to focus on higher-order writing skills]]>
Copyrights © 2026
