Human Vs AI-Supported Feedback: Effects On Academic Achievement, Self-Regulation, and Feedback Literacy*

Abstract

This study aimed to compare the effects of teacher feedback (TF) and AI-supported feedback (AIF) on academic achievement, perceived self-regulation, and feedback literacy among 42 sixth-grade students in a private school in Istanbul, Türkiye. Forty-two students were assigned to either a TF group (n=21), which received written feedback from the teacher, or an AIF group (n=21), which received AI-generated feedback through a Python-based natural language processing platform integrated with Cognitive Diagnostic Modelling. Both groups completed weekly quizzes over a four-week intervention period, aligned with English curriculum learning objectives. A 2 (time: pre-test vs. post-test) × 2 (group: TF vs. AIF) mixed-design multivariate analysis of variance (Mixed MANOVA) revealed significant improvements in all measured outcomes from pre-test to post-test (p