Skip to main content

When GenAI enters the assessment relationship: Public expectations of transparency, judgement and responsibility

PDF

Abstract

Human-centred learning design requires more than using generative artificial intelligence (GenAI) to make assessment feedback faster, more accessible or scalable. It requires attention to whether human judgement, relationships and responsibility remain visible when feedback is produced. The public is a key stakeholder in these decisions because university qualifications operate as public credentials whose legitimacy depends on confidence in how achievement is judged and certified. Yet debate has largely focused on students’ use of GenAI, with less attention given to public expectations of academic practice when GenAI contributes to feedback. This exploratory study surveyed 299 participants, most of whom had current or previous experience of higher education, to examine the implied obligations underpinning university assessment and feedback. Participants strongly endorsed feedback as an expected component of university teaching and generally supported disclosure of GenAI use, meaningful academic involvement and visible instructor responsibility. Exploratory factor analysis identified two provisional dimensions: AI feedback transparency and legitimacy, and assessment expectations and feedback value, although the second demonstrated only modest internal consistency. Interpreted through psychological contract theory, the findings suggest that human-centred GenAI implementation depends not only on feedback quality or efficiency but also on preserving visible academic judgement, accountability and educational purpose. The central design distinction is therefore between GenAI that supports academic practice and GenAI that substitutes for meaningful academic engagement.

Keywords

Generative AI, assessment, feedback, psychological contract, higher education, survey research.

Author Biographies

  • Dr Kate Lafferty, La Trobe University

    Dr Kate Lafferty is a Senior Lecturer in Assessment and Pedagogy at La Trobe University. Her research interests include assessment literacy, competency-based and inclusive assessment, micro-credentialing and the assessment of general capabilities. Drawing on more than two decades of teaching across primary, secondary and tertiary education, her work combines a strong research base with a practical understanding of assessment in educational settings.

  • Dr Nicole Brownlie, University of Southern Queensland

    Dr Nicole Brownlie is a Lecturer in Education at the University of Southern Queensland. Her research spans assessment, teacher wellbeing and the ethical use of artificial intelligence in education. Through her work on ethical augmentation, she examines how AI can support educational practice while preserving professional judgement, responsibility and meaningful human relationships. Her recent work has focused particularly on assessment, feedback and trust in AI-influenced education.

  • Dr Kellie Buckley-Walker, University of Wollongong

    Dr Kellie Buckley-Walker is a Senior Lecturer in the School of Education at the University of Wollongong. Her research focuses on classroom assessment, assessment and data literacy, and the validity of assessment evidence. She is particularly interested in how teachers design assessment tasks and analyse and interpret the evidence they produce, informed by her experience as both a classroom teacher and teacher educator.