How EFL Learners View Speaking Feedback: Artificial Intelligence vs. Human Instructors

Abstract

This research examines the attitudes of Iraqi undergraduate students majoring in English as a Foreign Language (EFL) toward the shift from reliance upon instructor-provided feedback to reliance upon generative artificial intelligence (AI) for receiving speaking course feedback. With the expanding accessibility of generative AI tools, it is important to understand how learners may view this transition in order to develop effective instructional practices for teaching second language (L2) skills. To assess senior students' perceptions, descriptive survey methodology was utilized to administer a structured questionnaire to 53 senior students enrolled at the University of Kufa, College of Education. Results indicate that respondents perceive both teacher and AI-generated feedback positively, with a slight preference for receiving feedback from an instructor (M=3.96) compared to receiving AI-generated feedback (M=3.83). Respondents view teacher feedback to have the greatest motivating effect while also appreciating AI-generated feedback as providing immediate feedback and helping them gain confidence to engage in low-anxiety practice (e.g., through immediate feedback). Finally, this research demonstrates a positive relationship (r=0.62) between students' perceptions of teacher-generated and AI-generated feedback, indicating that those students who value receiving feedback via one source also perceive value from receiving feedback from the other source. The research, therefore, proposes a hybrid feedback model that incorporates AI for repetitive drills but maintains the role of teachers in providing emotional and nuanced support, offering a practical way to improve speaking results in an overcrowded Iraqi classroom.