Bristol Institute for Learning and Teaching

I recently had the opportunity to facilitate three group discussions with students from across the University on their experience of group work and AI as part of the AI/Inclusive Assessment (AI/IA) sessions run by our Student Fellows Pratham Gupta and Gaurav Saxena. 

We’ll follow up with more detailed accounts of these events over the coming weeks, but I thought I’d take a moment to share the main three points from my discussions with students. 

#1: Students are using AI; regardless of whether they’re allowed to or not. 

Shock! Horror! Students – across every faculty and at every level of study – are using AI for both individual and group assignments. 

Around 50% of the students in the AI/IA sessions reported using it for ‘research’ – to summarise papers, to show them resources they may not be aware of and, specifically as a group, to help them come to conclusions or provide direction. Only one student I spoke to (of around 40) admitted to using it for more than those purposes. Students were much more negative about it being used for group tasks than they were about using it for individual assignments. 

I started each of my discussions with the same question; ‘what experiences have you had with AI and group work?’ and the same answer was given every single time: 

‘Someone in my group had very obviously used (generative) AI for their contribution but I didn’t feel like I could call them out on it.’

What happened next varied: when the group were getting marked as a whole, many of the students I spoke to said they had redone the work as they were worried about academic integrity issues.

When they were getting marked individually for their contributions, they just didn’t say anything in the hope that the tutor would notice and wouldn’t penalise the others. No one I spoke to had ever accused anyone else of using AI, even though they were confident they had encountered it.

#2: Feelings about AI use vary hugely depending on what faculty you’re based in. 

Students from the Faculty of Health and Life Sciences were the most negative about generative AI use – with the vast majority of students finding it to be a hugely unethical and problematic intrusion in their university experience. They viewed those who used it as ‘struggling’ and believed that they were only using it because they felt they didn’t have another choice. They also raised concerns around the lack of equity of access, with some of their peers having premium Chat GPT subscriptions and knowing others that were not able to afford it. 

Arts, Law and Social Sciences students were more mixed – with some feeling that it should be integrated more fully into the curriculum as ‘it isn’t going away’ but with others saying echoing similar concerns to their peers in Health and Life Sciences and wanted it separated from University as much as possible. 

Students from the Science and Engineering were almost exclusively positive about AI and were undoubtedly the keenest for it to be a part of their academic work. 

As one Physics student put to me: ‘we [as Science and Engineering students] are the most at-risk of AI taking over our jobs so it’s really important we know how to use it and how to use it well so we can show where our additional value can be’. 

#3: There isn’t a huge amount of clarity on how AI should be being used for group work assignments

Each of the sessions started with a panel discussion around the University position on AI and how it aligns with inclusive assessment – led by Prof. Tansy Jessop, Sarah Davies (Director of BILT and Education Innovation), the Faculty’s Academic Director for Programmes, Curriculum and Assessment (Dr. Dawn Davies, Dr. Cameron Hall and Prof. Bradley Stephens) and Dr Aisling Tierney (BILT Lecturer). 

The panel were encouraging of AI use on the whole and saw a future where it was increasingly integrated within the curriculum. Tansy discussed the four categories of AI use and a decent proportion (around 60%) of students across the 3 AI/IA events were aware of them. 

However, none of the students I spoke to had ever had the categories assigned to their group work projects; only individual assignments. Combined with the already tension-filled atmosphere group work can bring, adding in this layer of uncertainty around AI use is undoubtedly creating more anxiety for students. 

Conclusion

It’s clear there a lot of negative feelings associated with AI use – both in that some students are vehemently against its use, and also not feeling like it can be discussed when you believe someone has used it in work you are submitting as a group. Some students have strong ethical reservations about using it, whereas others feel it should be fully integrated if they are to be prepared to use it in the future. 

Where students are positive (largely in the Faculty of Science and Engineering!), they still would like to see more clarity over it’s accepted use and more openness in discussions with their peers about it, as currently there is a lot of secrecy and ‘shame’ around using generative AI. 

There are some practical implications that can be taken away from these sessions: 

1. As the lecturer, you must be clear in what is acceptable AI use for the assignment. Give examples of what is okay and what isn’t when setting the work. 
2. Consider how you are awarding grades for group work and what role AI may have in keeping this a fair and inclusive process. If students should use AI, do they have equal access to it? Equal skill levels? If they shouldn’t use it, how can you ensure that they haven’t? 
3. Think about creating a resource to support your students when setting up group work, which includes an element about discussing AI and the group’s position on its use. Study Skills has resources on group work but they are not mandatory for all students to take. A group work ‘cheat sheet’ or asking them to set ground rules may help set your group up for smoother discussions and help navigate difficult situations.