Keyword: automated feedback

ABSTRACT: Collaboration between humans and artificial intelligence (AI) has the power to transform education, but research has yet to fully address how students engage cognitively with AI-powered feedback. Although studies suggest that AI can improve writing, few explore how students perceive and interact with AI tools. To fill this gap, the present study investigated Chinese university students’ perceptions and experiences of using AI-powered English writing feedback tools: automated writing evaluation, generative AI, and corpora. Two hundred and ten student reflective journals were subjected to qualitative thematic analysis using NVivo software, along with analysis of classroom observations and students’ writing. The students evaluated AI-powered feedback tools in three dimensions: content quality, delivery method, and overall effectiveness. They felt that these tools provided better grammar correction, instant feedback delivery, and an enhanced user experience, but challenges included vague explanations, limited emotional connection, and risks of overreliance. Based on these insights, this study introduces the Student-Teacher-AI Collaboration Model for feedback writing, which is designed to enhance collaboration between students, teachers, and AI in foreign language education. The findings have practical implications for the integration of AI tools into writing instruction and will inform policymaking in the rapidly evolving educational field.

ABSTRACT: Automated writing evaluation systems are formative assessment systems that provide immediate, automated feedback on L1, L2, and EFL students’ writing in the form of writing quality scores and suggestions for revising. As such, these systems have the potential for alleviating some of the persistent barriers teachers face to implementing evidence-based writing instruction practices. However, simply adopting this technology without careful attention to how it is implemented will not guarantee instructional benefits. In this article, we draw on prior research to make recommendations to effectively integrate automated writing evaluation alongside evidence-based writing instruction practices to improve writing instruction and intervention, leveraging the affordances of this technology while addressing its limitations. Specifically, we discuss how researchers, interventionists, and educators using automated writing evaluation should develop students’ knowledge of underlying evaluation criteria; teach strategies for planning, drafting, and revising; supplement automated feedback with effective teacher-provided feedback; and enact goal setting and progress monitoring.