Metacognitive Reflection Loops in Human–AI Co-Writing: A Sequential Model for Strengthening Critical Reasoning Among Undergraduate Learners

Abstract

The rapid integration of generative artificial intelligence (AI) into higher education has transformed academic writing practices, creating opportunities for collaborative knowledge construction between students and intelligent systems. While AI-assisted writing tools enhance productivity, concerns remain regarding students' overreliance on AI-generated content and the potential decline of critical reasoning abilities. This study proposes a novel pedagogical framework, termed the Metacognitive Reflection Loop (MRL), which conceptualizes human–AI co-writing as a cyclical process of planning, AI interaction, reflective evaluation, revision, and cognitive reconstruction. The study aims to examine how sequential reflective engagement with AI-generated content influences undergraduate students' critical reasoning, metacognitive awareness, and academic writing quality. Drawing upon metacognitive learning theory, self-regulated learning, and cognitive apprenticeship, the proposed framework emphasizes deliberate reflection after every AI interaction rather than passive acceptance of generated text. A quantitative research design is proposed involving undergraduate students from multidisciplinary academic programs. Structural Equation Modeling (SEM) is suggested to evaluate the relationships among AI collaboration, reflective practice, metacognitive regulation, and critical reasoning. The expected findings indicate that structured reflection mediates the relationship between AI-supported writing and higher-order thinking skills. The study contributes to educational technology literature by introducing a sequential human–AI collaboration model that prioritizes cognitive development over technological dependence. Practical implications suggest integrating reflective checkpoints within AI-supported writing activities to foster independent analytical thinking in higher education.