What students think about AI and their learning
Artificial intelligence is a mixed bag for higher education. Last year, we were seeing the impacts of students experimenting with its application in their studies. While tech bros promised an AI utopia to investors, the realities of its use for plagiarism caught the attention of the press, policymakers and our institutions. In this quickly moving landscape, one voice was muted – that of the students. Peter Peasey (DEO) and I set about to fix that oversight and engage with our students through focus groups.
The focus groups were carefully designed not to influence students’ views so we could really see what matters to them (rather than telling them what matters to us and asking their opinion). We have written this all up in a peer-reviewed article so you can deep dive our methods, the context in the literature, and the results. For this blog post, I will just share some highlights.
Demographics
Student demographics are skewed towards Arts (19%), Social Sciences and Law (28%) and Life Sciences (36%) with only a small number of students from Science (10%) and Engineering (6%). The majority of students are post-graduate taught (38%) or post-graduate research (23%) while undergraduate representation is strongest for third year undergraduates (18%). Most participants are international students (62%).
Positives and negatives
Students found lots of reasons to love AI. Top of their list for positive benefits of AI was support for independent study skills with 126 mentions across the dataset, compared to the next more popular matter “efficiency gains” with 46 mentions. Other popular topics were support for writing, perceived improved quality of the learning experience, and support in understanding academic reading.
Just like the academic community, students saw that the major issue for AI is around academic integrity (77 mentions). They noted other major concerns as misinformation, equity issues (largely relating to premium services versus free services), worries about how using AI will lead to surface-learning rather than deep learning, and a decrease in higher order thinking and creativity.
Observations
Peter and I also made lots of observations while students were in discussion, as they flitted across topics in unexpected ways. For many, the definition of what is or isn’t AI is still unclear. A big cause of stress was what students understood about how AI content is detected by Turnitin. Some students who never even used AI had a worry that they could be accused of cheating. Students wanted to know what academics could see via these detection tools so they could understand that process more and redress some anxiety.
Many students considered it farcical that some Universities (such as in the USA) have outright bans on AI. As one put it “The toothpaste is well out of the tube. It’s absurd to be told not to use it the same week we were taught about it”.
Some students also felt emergent existential dread because of AI, asking what is the point of HE given what AI can achieve and replace. This dread includes questioning the virtue of degrees, the merits of HE in general, and the ultimate utility of humankind. Heavy stuff indeed.
I was most surprised to learn that differences in opinions (and indeed morality) on the use of AI is causing friction between cohorts and amongst friends on different degree programmes. In such cases, unease, confusion, differential guidance, and different perspectives on the use of AI are enflaming friction within the learning community.
Many showed concern for how assessments might change to include more exams which caused anxiety for some while a small group saw it as an essential step. Students were emotionally charged when talking about the idea of AI marking their work. Students responding to “How would you feel if your work was marked – in part or in full – by AI?” use words such as horrible, outraged, uncomfortable, worried, disappointed, annoyed, unwilling and unhappy.
Most students agreed that it was perfectly sensible that different subjects would engage with AI differently.
Numbers
At the end of the focus groups, we asked students to fill out a short form, which gave us some nice clean numbers to go with the qualitative data collection methods.
- 79% already used AI to support their students.
- 56% plan to use AI to support their studies in the future (36% unsure; 8% no).
- 54/67 students used ChatGPT.
- 50% thought that AI has the potential support learners under themes of accessibility, widening participation, and equality, diversity and inclusion. (30% maybe; 20% no).
- 86% worry that AI will impact academic integrity in assessments.
- 63% believe there is an appropriate way to incorporate AI into their learning experience (28% unsure; 9% opposed).
- 80% are opposed to AI marking their work, aside from MCQs.
- 53% believe that AI risks the perceived value of degree programmes (35% unsure; 12% no).
- 62% believe that AI skills are important for future employers and careers (27% maybe; 11% no).
Discussion and recommendations
The focus group data was thematically analysed into various categories of concern, and then mapped to categories drawn from the literature. The main result of this is that students and staff are pretty much on the same page with shared concerns around AI. Very reassuring!
We take time in the paper to think about what all this means for teaching and learning in the future. We discuss plagiarism of course. But we also delve into what this means for curriculum change. Authenticity is a highlight here and really reflects what we are doing with our assessment and feedback strategy at Bristol where we think about authentic assessment, how we can be inclusive and design for all, and how we can take a holistic and integrated approach. The list of recommendations includes lots of things that have already been done at Bristol, such as sector-learning training resources for students and putting resource behind AI & Assessment matters (happening via BILT-funded projects).
The academic paper: https://doi.org/10.58459/rptel.2025.20005
Recommended next step – check out JISC’s Assessment Ideas for an AI-Enable World. It’s a fantastic resource!
This blog forms part of our ‘A – Z of BILT’ blog series – look out for ‘B is for… ‘ next week! View the full series here.
