Preserving Cognitive Ownership in Higher Education: A Sustainable Hybrid Pedagogical Framework for Reasoning-Centred AI Integration

This study explores how distinctive types of generative artificial intelligence (AI) practice, augmentation, co-construction, and replacement shape students’ reasoning skills and sense of cognitive ownership in higher education (HE) academic writing. This research also responds to growing humanitarian concerns about the erosion of student commitment, the undermining of autonomy, and ethical learning in HE. To address this core gap, an explanatory sequential mixed-methods design was employed. Data were collected from 412 UK HE students, complemented with in-depth interviews from 24 participants. Quantitative modelling showed that augmentation strengthens reasoning through reflective engagement, co-construction yields mixed cognitive outcomes, and replacement significantly weakens ownership and efficacy. Qualitative findings revealed subsistent experiences behind these practices: some students articulated no ethical harm by AI-supported reflection, while others exhibited a quiet disarticulation of their self-learning skills. Incorporating these insights, this study proposed the Hybrid Human–AI Reasoning Integrity Model (HHARIM), a sustainable pedagogical framework in HE that centres human reasoning in ethical AI use. The recommended model also highlights cognitive ownership as an essential element and outlines a robust framework for responsible AI use to safeguard learning, ethics, and autonomy in HE. This study contributes theoretically by offering HHARIM as a framework for effectively embedding AI, thereby upholding ethical, sustainable, and human-centred learning. Ultimately, the implications of this proposed model will influence HE systems to encourage sustainable AI pedagogical practices that reinforce academic writing rather than compromise students’ learning efficacy.