Who Uses AI, How, and When?
By Matthew Voice, Department of Applied Linguistics, University of Warwick
As I mentioned in my previous WJETT blog post, I have participated in a strand of the ‘AI in Education’ learning circle during the last few months. Over the course of our strand’s meetings regarding the role of AI in formative feedback, our focus has primarily been on the nature of the emerging generative AI technology. Our research and conversation has a great deal to contribute with regards to what large language models mean for the future of formative feedback. However, given that these models are tools, it is worth reflecting on the user, too.
Our learning circle represent educators and students from a wide range of degree programmes, and motivations for engaging with higher education will vary from student to student across all courses. Throughout our discussions, conversation has largely focused on the role of AI in the improvement of formative feedback for students who are motivated by a desire for academic excellence. Given that students in this position will likely be motivated to engage with extracurricular activities within the university (e.g. joining the WIHEA as a student member, partaking in surveys of student experience or voluntary training workshops), their voices will perhaps be heard most clearly during our present conversations.
But the experiences of these students are not representative of our cohort as a whole, or indeed of all points across an individual’s university journey. Many students may view academic writing – either across their degrees, or for certain assignments or modules – as an obstacle to be overcome. In these cases, the objective of academic writing shifts away from excellence, and towards the production of work which will simply allow the student to attain a passing grade. Engagement with GenAI for these students might not mean refinement and improvement; it may simply be a resource for the fastest or most convenient means of content generation.
Integrating GenAI into teaching and formative feedback requires a recognition of this spectrum of experience and motivation. In my previous WJETT blog I recommended that future discussion and planning should consider the reasonableness and realism when supporting students to use and think critically about GenAI. By this, I mean:
1) Reasonableness: Is the task being asked of the GenAI (e.g. to produce formative feedback on a draft assignment) one that it is capable of achieving?
2) Realism: Does our (educators, policy planners, researchers) understanding of engagement with GenAI reflect actual independent student use cases?
Assessing reasonableness through understanding what GenAI is capable of achieving will require ongoing review, and the development of support and training for staff in order to keep pace with developments. This, I think, has largely been the focus of the work done by the ‘AI in Education’ learning circle and the final report this group has produced. Going forward, we also need to consider how well we understand our students’ independent engagement with this sort of assistive technology. What tools and resources are they familiar with? How do they understand them? What do they do with them, and at what point(s) in their academic writing?
Grounding future policy and pedagogic resource development in relation to a realistic model of students’ use and understanding of GenAI will be a task as complex as anticipating the impact of future technological development in large language models. By acknowledging this, and undertaking this work, we best position ourselves to ensure that outputs and resources which arise from working groups such as ours will be meaningful to staff and students working across the university.