Review finds varied student preferences for AI and teacher feedback

Updated Sep 10, 2026

What the study reports

Yağmur Öztürk and Ayça Çebi review higher education students’ perspectives on AI-generated feedback in online learning. They consider cognitive and emotional benefits, preferences relative to teachers, and expectations of feedback.

Preferences varied: some students favoured teachers, some AI, and many a combination. Students wanted detailed, constructive responses suited to context. These preferences do not establish that a combined approach is best or demonstrate learning gains.

Questions for a local feedback pilot

Our suggestion is to separate students’ preferences from the quality and consequences of feedback. Ask whether a comment is understandable and useful, then examine whether it is accurate against the actual task. Liking a tool and benefiting from it are different outcomes.

Define staff responsibility and opportunities for follow-up before a pilot begins. Compare the intended use with alternatives appropriate to the course, and record who did not participate. Our governance checklist can support this planning. These are editorial proposals, not tested benefits of an analysis product.

Reference

Öztürk, Y. and Çebi, A. (2026; online 22 November 2025). The potential of AI-generated feedback from the students’ perspective: a systematic review. Assessment & Evaluation in Higher Education, 51(1), 41–62. Paper.