Rethinking assessment in the face of AI – example of an in-year change

Since the launch of open-access Generative AI tools at the end of 2022, educators across the globe have had to stop and question their assessment methods, responding to the potential threat of academic misconduct from these tools. Whilst it is not realistic or advisable to simply return to in-person invigilated exams or to ask every educator to significantly rewrite every assessment, it is possible to make changes to assessment in-year that reduce the likelihood of AI-related misconduct, and that increase student engagement with assessment as learning.  

Sian Robinson, Programme Director for BSc Business and Management and BSc Business Studies in the University of Exeter Business School, spoke to us about in-year changes she has made to the MSc HRM module ‘Employment Relations’ (BEMM045) which ran in Term 1 of 2023/24. Following a workshop during the EduExe Festival in summer 2023 on AI and Assessment, Sian decided to adapt her module’s assessment to make it less susceptible to AI misconduct, whilst also embedding greater support to enhance students’ critical evaluation in their writing.  

What’s changed? 

 Using the same module descriptor, Sian has kept the ‘Individual written assignment’ (100% of module grade), but has replaced a 3000-word essay with a two-part task that embeds reflection on learning and reduces the possibility of students turning to Generative AI for assistance.  

Part A – Review of weekly discussion activities (1500 words) 

Students choose three topics from the course, and write 500 words in response to a series of short tasks, including:  

– summarising key discussion points from the lecture 

– reflecting on their contributions to the discussions 

– providing a list of related sources (with a reflection on their learning from these) 

– providing an organisational example of this topic  

– analysing the importance of the topic within the module context  

 Part B – Critical evaluation essay (1500 words) 

Developing their learning from one of the three chosen topics for Part A, students write a critical evaluation of the importance of their topic for the employment relationship, including critical engagement with theory, HR practice and examples, and relevant literature.  

What’s the impact? 

 Sian notes that the updated assessment task is deliberately designed to increase student engagement with weekly activities in class, and that this has impacted the quality of discussions during the course, with positive feedback from students. Although this required some careful thought and additional planning to ensure that session activities were structured in such a way that students would be able to engage during class and reflect on their learning afterwards, Sian observed a shift of focus onto term-long learning and weekly reflection that helped to embed critical thinking and engagement with relevant literature and examples that students could draw on for the extended Part B of their assignment. The change in assessment model also required Sian to update her approach to providing formative feedback, adding discussion of an exemplar ‘Part A’ of the submission to other support from previous years, including individual 1-1 sessions and an in-class activity where students worked with exemplar submissions, using the marking criteria and group discussions to evaluate each piece and suggest improvements.  

Overall, Sian feels that the new assessment task demonstrates stronger constructive alignment with the module ILOs, alongside scaffolded support for students’ learning. Whilst marking the submissions, Sian has noticed that the assignment’s increased focus on reflection more clearly highlights submissions where students have genuinely engaged with the course content and developed the relevant skills and knowledge being assessed. The emphasis placed on reflection has helped reduce the impact of Generative AI, as these tools cannot convincingly produce text that covers all of the reflective elements set out in Part A of the module assignment, and would be less able to show meaningful development of these in the extended critical evaluation of Part B.  

What’s next?  

Looking ahead, Sian is planning to review Part A of the assignment, perhaps reducing the number of topics that students write about so that the word count encourages greater depth of reflection on learning. Elsewhere, as Programme Director, Sian hopes to share her practice with colleagues to show the way in which this approach could be adapted elsewhere to increase student engagement, embed greater reflection on learning, and reduce the potential impact of Generative AI. 

For more information please contact: