Artificial intelligence as a catalyst for transformation: EFL learners’ perceptions of AI-powered language tools in delivering effective oral and written corrective feedback

Abstract

The integration of artificial intelligence (AI) in English as a Foreign Language (EFL) teaching and learning has reshaped corrective feedback (CF) practices, facilitating new opportunities for both oral and written language development. This study examined EFL learners’ perceptions of AI-powered CF, focusing on the effectiveness of six types of oral corrective feedback (OCF), namely clarification request, elicitation, explicit correction, metalinguistic feedback, recast, and repetition, and three types of written corrective feedback (WCF), including direct, indirect, and metalinguistic feedback delivered through the AI tools Kippy and Pi. Adopting an ethno-phenomenological design, the study collected the data through a Likert-scale questionnaire and semi-structured interviews with 27 Iranian EFL learners at the B1 proficiency level. The results revealed that explicit correction and direct feedback were rated as the most effective CF types, while repetition and indirect feedback were perceived as the least helpful. The interview data supported these results, highlighting the learners’ preferences for direct and immediate feedback that clearly signals errors and provides opportunities for correction. Participants reported that AI-generated CF enhanced their autonomy in self-correction, created space for language learning, and delivered instant, personalized feedback tailored to their individual language needs. However, they noted that AI tools prioritized linguistic accuracy over communicative meaning, requiring teacher intervention for deeper contextual understanding. This study offers a comprehensive understanding of how AI tools may support CF processes, contributing to the integration of AI into EFL pedagogy.