Human vs. AI Feedback on EFL Students’ Essays: A Comparative Study

Abstract

This study compares the effectiveness of human and AI-generated feedback on improving EFL student essays. Twenty students from a state university in Türkiye participated, producing 19 argumentative and 20 opinion essays. Initial scores ranged from 2.00 to 2.15. Essays were scored with a four‑band analytic rubric adapted from Reid (1993) and expanded with grammar‑error categories from Schenck & Daly (2012). Human feedback significantly improved essay quality by addressing grammar, structure, and content with personalized comments, encouraging deeper engagement. In contrast, AI feedback focused mainly on grammatical corrections and clarity, lacking comprehensive and personal touch. The findings suggest a hybrid approach, utilizing AI for basic corrections and human instructors for detailed feedback, which could optimize student learning outcomes.