Abstract:

The rapid expansion of Generative Artificial Intelligence (GenAI) has created both opportunities and challenges for Higher Education (HE), particularly in relation to academic integrity, authorship, and student learning. This pilot study examined two emerging AI supported tools designed to promote responsible AI use, strengthen authorship authenticity, and support students’ comprehension and writing development. Using a mixed methods approach across two university cohorts at a leading University in Australia, the study explored students’ experiences of AI tools offering personalised authorship quizzing (Auth+) and AI assisted concept explanation from video content (Tutbob). Findings indicate that students generally viewed Auth+ as a legitimate mechanism for checking authorship and reflecting on their writing, though high failure rates and cognitive load concerns highlight the need for caution when interpreting quiz performance. Tutbob was perceived as highly useful for simplifying complex material, enhancing understanding, and improving academic writing, despite technical limitations in processing fast or noisy audio. Across both tools, students expressed nuanced perspectives on AI, recognising its benefits while emphasising risks of over reliance, unfair advantage, and diminished critical thinking.

The results reinforce emerging international literature arguing that AI solutions must be embedded within broader institutional strategies rather than used as standalone detection mechanisms. The study highlights the importance of assessment redesign, AI literacy, transparent guidelines on permissible AI use, and inclusive tool design aligned with Universal Design for Learning (UDL). Together, these findings demonstrate that AI tools can support learning and academic integrity when deployed ethically, transparently, and in alignment with supportive institutional policy frameworks.

The study reinforces growing international agreement that reliance on post-hoc AI detection is neither reliable nor equitable. Institutions should prioritise assessment formats that foreground process, authenticity, and performance, including staged submissions, oral components, in class creation tasks, and multimodal assessments. Authorship testing can support this shift by encouraging deeper authorship engagement, but only as one piece of a larger integrity strategy.

Given the potential for misclassification, bias, or unintended integrity risks, student-facing AI tools require robust oversight. This includes:
- Transparency about how tools work and what data they use,
- Opportunities for students to contest results (particularly with authorship quizzing),
- Regular review by academic integrity and learning design experts, and
- Institutional accountability for ethical deployment.

These implications highlight the need for universities to adopt a coordinated, institution wide approach to AI that integrates pedagogy, policy, and ethical governance to ensure that generative AI strengthens, rather than compromises, the quality, fairness, and integrity of HE.

Keywords:

AI, authorship, learning.