Authentic assessment is essential

Authors: Emeritus Professor Jim Mienczakowski and Emeritus Professor Greg Whateley

July 2025

In recent weeks you may have noticed that the media in Australia has been stepping away from stories relating to AI plagiarism, student cheating and non-student authored assignments. The AI pollution of research data, however, is still on the agenda.

The lack of media interest in the veracity and integrity of educational assessment processes is concerning. It’s not as if institutions have resolved their assessment worries around ‘Chat GPT’ written student assignments – many have simply offered further elaborate student guidance on AI use or softened their parameters around the permissible use of AI. It seems that 2025 is destined to become ‘the year that Mr. GPT did my degree’ for some (many?) students.

And will we ever know how much AI penned material is actually being submitted by students in their academic submissions?

The current answer is ‘no’.

And in order to really identify Chat GPT’s contributions to any assignment you need not only fail-safe mechanisms for detecting its use but also the will and resources to deal with what you find. There are complex demands at play in respect to corralling AI use by students.

Back in the early days of (non-technological) contract cheating, some universities were faced with hard choices to make. Identifying cohorts and cliques who’d used particular contract services and dismissing all involved or counselling guilty students that cheating was wrong and that they needed to work differently. Great care was generally given to ensure that international students, in particular, were careful not to cheat again but also that they were not broadly dissuaded from continuing their studies in Australia. International students are, after all, the full fee-paying students that universities have become so dependent upon for their fiscal well-being.

A decade later

Jump ahead a decade to today’s tertiary environment and the ‘too hard basket’ is full of partial answers, weak solutions and fear of market share losses if one university completely closes the door on Chat GPT written assignments whilst others (around the globe) take a less adversarial (non-punitive) position and are ‘ambivalent’ in terms of AI’s writing of students’ assessment pieces.

For all of the talk of academic culture, standards and learning outcomes our universities are still huge corporations which increasingly rely upon lower paid adjunct staff, markers and teachers to be the blunt end of the assessment system. They do not want to spend more money on assessment processes unless they have to.

To those at the brunt end of the assessment scenario dealing with large numbers of student assignments - AI incursions must be a nightmare. Where once academic misconduct by students was not that common – the ease of use of GPT style bots has leant cheating an instant utility that contract cheating and plagiarism lacked.

Partial acceptance of AI

The partial acceptance of AI as an aid and tool for student learning might also make the gap between doing background organization and even research via AI and having a bit of GPT help in writing an assignment seem even more permissible to students. The more it goes undetected in its use – the more students will be tempted to try their luck and use it inappropriately.

The ultimate test of it all is, of course, to have an ultimate test for all students. Invigilated exams and Viva Voce assessments. Herein, universities can (despite the unpopularity of examination style assessment in some disciplines and even in some States such as Queensland) ensure that students know what they should know and can demonstrate their own learning.

More than ever, invigilatorPlus style applications, bringing together the strengths of AI monitoring and human invigilators, should be seen as key strategies for universities to restore confidence in their ability to protect the integrity of their student learning and achievement outcomes.

Without vigilant integrity approaches, university degrees are destined to gradually ebb in their perceived value whilst AI continues to relentlessly develop its people-replacing potentials.

Emeritus Professor Jim Mienczakowski is a Higher Education Consultant based in Melbourne.

Emeritus Professor Greg Whateley is currently CEO and ED at the Australian Guild of Education (Melbourne).