The recent rise of accessible generative artificial intelligence (GenAI) platforms has changed the way we harness and produce information. The emergence and growth of GenAI poses challenges for educators to support students to use this technology responsibly and with academic integrity. Published evidence verifies that educators demonstrate low levels of accuracy in distinguishing between student and AI generated writing. Additionally, current works in AI detection outline mixed results in efficacy, further promoting a multi-pronged approach to dealing with academic integrity related issues with AI. However, GenAI also presents a wealth of opportunities for efficient and effective learning and teaching. As student numbers within many higher education institutions increase and staff are inevitably having to do more with less, GenAI can facilitate personalisation of learner experiences and provide tailored and reflective assessment feedback. Hence, this study analyses Auth+, an innovative artificial intelligence tool, designed to create, deliver and monitor personalised quizzes for students on their own assessment submission. The emphasis of Auth+ is to use emerging technologies to enable educators to take a novel approach to academic integrity by cross-questioning students on their own submissions at scale. The premise of Auth+ is that writing is thinking. As a result, writing is understood as a sequential experience that builds episodic memories, which can be argued is the foundation of learning. With the growth of generative AI, educators need to protect students from the risks of outsourcing their learning as a potential route to disconnection from the learning experience and ultimately skill shortages and devalued qualifications. Auth+, hence, generates short timed ‘authorship tests’, which are designed to diagnose the presence (or lack thereof) of these episodic memories, or self-conversations, formed through drafting, reviewing and revising a summative piece of work. The emphasis is not on AI detection, but on fostering students’ writing skills by promoting self-conversation during the process of writing. Auth+ is designed to support and develop students’ understanding of academic integrity and foster productive and responsible approaches to assessment completion. Through in class trials, we investigate the utility of this AI based personalised ‘cognitive’ quizzing and explore students’ attitudes and responses towards this approach to authorship detection. Through this study, we gain and share insights on the potential value of using AI generated authorship quizzes both to students and to educators working within higher education.

During the trial, students engaged with the Auth+ platform during two in class mini quizzes to assess their authorship of:
1) an AI generated draft; and
2) their own submitted assignment. Results were compared to explore the efficacy of Auth+.

Additionally, students’ experiences of using Auth+ were sampled to reveal its impact on student satisfaction and awareness of issues of academic integrity. The results are discussed in relation to the application of the platform for effective delivery of quality teaching and assessment for large cohorts. Additionally, we put forward the case that practical engagement with Auth+ can stimulate student awareness about the responsible use of GenAI.

Keywords: Artificial intelligence (AI), academic integrity, authorship, assessment.