<![CDATA[The rapid adoption of artificial intelligence (AI) in higher education has reshaped students’ academic practices, particularly in relation to academic writing. While prior studies have largely examined attitudes or perceived effectiveness of AI-based tools, limited attention has been given to how AI-supported writing unfolds as a learning process shaped by student agency, contextual constraints, and reflective practice. Addressing this gap, the present study adopts a qualitative, process-oriented perspective to explore how higher education students experience and perceive the use of ChatGPT in supporting academic writing. Using a qualitative research design, semi-structured interviews were conducted with nine higher education students representing language studies, social sciences, and exact sciences. Data were analysed thematically to capture patterns related to writing support, perceived learning processes, challenges, and students’ strategies for reflective and critical AI use. The findings indicate that ChatGPT supports academic writing by reducing initial barriers to writing, enhancing engagement and motivation, and providing personalised feedback that facilitates language development and revision. Students reported improvements in writing quality and increased confidence, suggesting that AI can function as both cognitive and affective scaffolding. However, these benefits were accompanied by concerns regarding rigid or inconsistent AI-generated responses, uneven technological competence, and limited institutional guidance. Importantly, students demonstrated critical awareness by evaluating AI outputs and combining them with other academic sources, highlighting reflective and responsible use rather than passive reliance. Overall, the study conceptualises AI-supported academic writing as a dynamic and mediated learning process rather than a discrete technological intervention. By proposing a process-oriented qualitative model, this study contributes to ongoing debates on AI in higher education and underscores the need for pedagogical strategies and institutional frameworks that promote ethical, reflective, and learning-oriented integration of AI technologies.]]>
