L2 writer engagement with automated written corrective feedback provided by ChatGPT: A mixed-method multiple case study

Abstract

Automated written corrective feedback (AWCF) has been widely applied in second language (L2) writing classrooms in the past few decades. Recently, the introduction of tools based on generative artificial intelligence (GAI) such as ChatGPT heralds groundbreaking changes in the conceptualization and practice of AWCF in L2 pedagogy. However, students’ engagement in such an interactive and intelligent learning environment remains unstudied. The present study aims to investigate L2 writers’ behavioral, cognitive, and affective engagement with ChatGPT as an AWCF provider for writing products. This mixed-method multiple case study explored four L2 writers’ behavioral, cognitive, and affective engagement with AWCF provided by ChatGPT. Bearing the conversational and generative mechanisms of ChatGPT in mind, data on students’ engagement were collected from various sources: prompt writing techniques, revision operations, utilization of metacognitive and cognitive strategies, and attitudinal responses to the feedback. The results indicated that: 1) behavioral engagement was related to their individual differences in language proficiencies and technological competencies; 2) the participants have failed to metacognitively regulate the learning processes in an effective manner; and 3) ChatGPT ushered in an affectively engaging, albeit competence-demanding and time-consuming, learning environment for L2 writers. The study delivers conceptual and pedagogical implications for educators and researchers poised to incorporate GAI-based technologies in language education.