The delivery of teaching and assessment in higher education has faced significant challenges over the last three years which have affected all those involved in the quality arrangements associated with the delivery of teaching, assessment, and the student experience. This blog is the first in a series which looks at how the UK sector leads on quality in higher education. For this first blog, Professor Clare Peddie, Vice Principal (Education) at the University of St Andrews and Deputy Chair of the UKSCQA, looks at the sector’s response to current challenges associated with the introduction of generative AI tools such as Chat GPT.
When looking at responses to generative AI tools I see a lot of knee-jerk reactions, from initial concerns in the media around the risks to the academic integrity of assessments to the rapid introduction of new online detection tools, as well as the return to in-person pen-and-paper examinations and changes to academic misconduct policies. Now, because of the many webinars, roundtables, and evidence-gathering conversations, thinking is developing around how we can respond positively and in a measured way. In the same way that the COVID-19 pandemic accelerated pedagogical advances in digital education, generative AI tools are, in my view, accelerators of the prevalence and quality of good pedagogical practice. These new tools will stimulate changes in pedagogy that we have been advancing throughout higher education over many years but will now quickly receive more prominence in our practice.
Developments in education which will be accelerated by the introduction of Generative AI include:
- the incorporation of experiential, personalised, and reflective learning practices;
- the advancement of activities designed to improve the metaskills of our graduates;
- developments to incorporate more authentic assessment, i.e., assessment aimed at the application of knowledge to real-life or work-based problems generating meaningful and worthwhile educational experiences; and
- the incorporation of project, group, and interdisciplinary activities.
In some institutions, including my own, the essay has been the mainstay of assessment in some of our more traditional disciplines. It would not be unusual for modules to be continuously assessed by essays and an examination asking further essay questions. Prior to Chat GPT, many educators and pedagogical researchers were already questioning the authenticity and the transferability of the skills graduates gain from these types of assignments. Now, we can predict that changes to the curriculum may involve greater involvement of the academic tutor in the development of students’ critical thinking skills, and discussion of the plans for the preparation and submission of an assignment.
Before the final summative submission, the tutor or lecturer may, for example, supervise the:
- development of a novel research question or reflective assignment;
- preparation of an outline plan;
- justification for selection of the resources that will be consulted; and/ or
- development of the argument and formative assessment of a draft document.
These measures will help preserve the academic integrity of the process while also supporting the educational purpose of assessment.
Some of our more innovative curriculum developments already in place will be largely unaffected by generative AI. For example, the Vertically Integrated Projects (VIPs) which are innovative interdisciplinary research-based projects that enable students from different year cohorts and academic staff to collaborate on long-term real-world issues which are part of the curriculum in number of institutions, including my own, and are largely immune to these risks to academic integrity.
Intensive personal supervision of the learning and assessment process may mean that the workload for academics will increase but, arguably, that the quality of the learning experience will be significantly improved. Consequently, curricula may have to change by substantially reducing the number of assessments. Again, many of us will recognise our tendency in the sector to ‘over-assess’, and we know that multiple assignments with feedback on which students have little time to reflect and learn from are less meaningful to a successful education. So, this is another way in which AI could further catalyse good pedagogical practice.
Like the digital transition catalysed by COVID, the transition to a new way of teaching will be pedagogically demanding. We find ourselves challenged, once again, to develop the support and guidance to help educators adapt, find the space in our commitments for curriculum development, and put in place the agile mechanisms by which assessment and teaching methodologies can be easily adapted. The question for me is not what we need to do to respond appropriately to the introduction of AI but how we resource that change and support the wellbeing of all those at the sharp end of curriculum delivery.
By Professor Clare Peddie, Vice Principal (Education) at the University of St Andrews and Deputy Chair of the UKSCQA.