The effectiveness of artificial intelligence-based personalized feedback in teaching English as a foreign language

Abstrak

This study is important because students' low writing skills indicate the need for faster, more accurate, and personalized feedback. This study used a quasi-experimental mixed-methods design involving 35 tenth-grade students selected through purposive sampling. Data were collected through analytical writing tests, questionnaires on learning interest and independence, observations, and interviews, using rubrics and standardized instruments. Quantitative results show that AI feedback has a significant positive impact (large effect size). After receiving AI feedback, students' writing performance scores increased by an average of 6.63 points (9.2%), interest in writing increased by an average of 20.1%, and independent learning skills increased by 15.8%. AI feedback is significantly superior to traditional teacher feedback in overall teaching effectiveness. Qualitative findings indicate that students appreciate the AI system's real-time responses, personalization features, and non-judgmental approach. This study concludes that personalized AI feedback is an effective pathway to transform writing instruction from teacher-centered to student-centered, and suggests adopting a human-machine collaborative model in teaching practice.