AI as a Personal Tutor: Enabling Students to Access Smart Tools

Daniel Lee (School of Education, Adelaide University, Adelaide, Australia)
Amy Robinson (School of Education, Adelaide University, Adelaide, Australia)

Journal of Research in Education and Pedagogy
Vol. 3 No. 3 (2026): Journal of Research in Education and Pedagogy

Generative Artificial Intelligence (GenAI) is increasingly used by university students, yet enabling/preparatory cohorts may require explicit guidance to realise learning benefits while maintaining academic integrity. This study explored GenAI tools as “personal tutors” within a leading Australian university preparatory context, focusing on how students actually used these tools, how institutional transparency requirements shaped practice, and what barriers remained. A blended descriptive methodology combined autoethnography of teaching practice with student-generated artefacts (Statements of AI Engagement and assessment extracts). Dataset 1 comprised the educator’s reflective journal notes, analysed using Reflexive Thematic Analysis. Dataset 2 comprised de-identified student artefacts, analysed using Inductive Thematic Analysis. To contextualise the qualitative findings, Statements of AI Engagement were also coded according to the institution’s four AI-engagement levels and to the tools and usage reported, and then summarised using descriptive frequencies/percentages. Participants were students (n=74) enrolled in two university preparatory courses (English Literacy and Research Design) during Semester 2, 2024. 72 Statements of AI Engagement were collected from 98 submitted assessments. Findings indicate broad acceptance of GenAI as supportive learning aids, most commonly for idea generation, rephrasing, summarisation, and basic editing/proofreading. ChatGPT was the most frequently acknowledged large language model, while AI-enabled search and writing-support tools were used more selectively. Although students generally positioned themselves as retaining authorial control, the study also identified persistent concerns regarding over-trust in automated outputs, cognitive load associated with evaluating and integrating AI suggestions, and an internalised stigma that reduced willingness to access both human and AI writing support. The findings suggest that structured, guidelines-based integration, including explicit verification instructions and reflective Statements of AI Engagement, can promote responsible, transparent GenAI use in enabling education.