How would you feel if your lecturer used AI to help draft your feedback? And would you feel differently about AI being involved in marking your work?
That’s what we asked Westminster students in January 2026. Our earlier survey of 393 students suggested people felt quite differently about AI helping with feedback and AI doing the marking, so we sat down with students to understand why.
We held four small focus groups across New Cavendish, Marylebone, Harrow and Regent, and two students gathered views from peers in Life Sciences and Business Studies.
The small-group format gave students the opportunity to discuss their views and experiences in more depth.
AI–supported feedback – benefits with boundaries
Students could see potential benefits, particularly around making feedback clearer and more consistent, while also saving tutors time and potentially helping students receive feedback sooner.
At the same time, students were clear that the tutor should remain central to the process. AI could support and build on a tutor’s feedback, but not replace their judgement. As one student explained:
“A human tutor should always have full oversight.”
A key concern was that AI-generated feedback could feel generic or miss the personal understanding of their work that a tutor brings. Students wanted feedback to remain personalised and reflective of the individual piece of work.
AI and marking – a more cautious response
We asked students whether it would matter what type of work was being marked. It did. Some students could see a potential role for AI in objective or standardised assessments but were much more cautious about subjective work such as essays or lab reports.
Students saw potential benefits, including easing pressure when a department is understaffed and helping tutors with the marking process. They also mentioned possible risks. One student worried that if many people wrote an assignment in a similar way, AI might mistakenly interpret this as plagiarism.
Focus group participants were clear that human oversight should remain central, with tutors retaining responsibility for final grading decisions. Students strongly opposed fully autonomous AI marking, pointing out its limitations in understanding context and academic nuance. As one emphasised:
“I would definitely be against fully AI-based marking”
And another reflected:
“The AI hasn’t been through the same training that the lecturer has gone through.”
What we are taking from this
Students aren’t necessarily asking us to keep AI out of the assessment process altogether. Instead, they are drawing an important distinction between AI supporting academic colleagues and AI replacing human expert judgement. Speed and consistency may be welcome, but human oversight and a personal understanding of students’ work remain important.
In one of our upcoming posts, we will widen the lens and look at where students think universities should draw the line with AI more broadly.
