• I Putu Eka Indrawan

  Author

• Dr. I Made Citra Wibawa

  Author

English education is on the verge of a transformation with the advent of Artificial Intelligence (AI) tools, which offer automated grading of writing and the provision of personalized learning assistance. But fully automated feedback is sometimes lacking in pedagogic judgment, and the reliability of feedback and meaningful teacher engagement are suboptimal. This study presents a Human–AI Collaborative Teaching Framework (HACTF), which combines AI-supported essay evaluation and a teacher-in-the-loop approach in adaptive teaching of English writing. The framework was tested with the IELTS Writing Dataset with 1435 essays with examiner comments and examiner band scores (Task Response, Coherence and Cohesion, Lexical Resource, and Grammatical Range and Accuracy) with the Overall Band Score. The essays were preprocessed and divided into 80% training (1,148 essays) and 20% testing (287 essays). Each essay was first scored by the AI, and then a confidence-based decision module decided whether teacher validation would be needed or not for the delivery of personalized recommendations. Experimental tests proved that the proposed approach provided 96.8% of feedback agreement and 97.6% of feedback acceptance and decreased the teacher intervention rate to 31.4%. Furthermore, the average feedback generation time has dropped to 4.8 minutes, the student writing improvement was 25.7%, and the overall gain of the students in IELTS band score was 0.86. The results suggest that AI efficiency can boost the quality of feedback, student writing ability, and the scalability, reliability, and individualization of writing instruction for the next generation of classrooms.

• PDF