Skip to main content Accessibility Statement

About this project

This project will test whether structured reflection with generative AI can improve student learning. Students will use a six-question reflection framework to explain ideas, make links between topics, apply knowledge and identify where they are unsure. The AI won’t be providing answers but will instead act as a questioning coach, helping students spot gaps, misunderstandings and weak reasoning. The study will compare learning, confidence and short-term retention across this approach, while also exploring how students experience AI in this role.

 

The project also aims to create practical, open-access resources, and staff development materials so the approach can be used more widely. Overall, it asks whether AI can support deeper, more independent learning when used to prompt thinking rather than deliver ready-made answers for students across higher education.

Reviewer discussion

This project is designed to investigate whether structured reflection supported by generative AI can strengthen student learning in higher education when the AI is used as a reflective guide rather than an answer generator. The project has five main objectives. It will examine whether AI-supported reflection improves students’ understanding of weekly topics, their short-term retention of learning, and their metacognitive calibration, which means how accurately they can judge what they do and don’t understand. It will also evaluate the quality of student reasoning through a reflection rubric, explore how they experience this form of AI-supported learning, develop a framework that can be applied across different disciplines, and produce practical resources to support wider use in teaching.

 

The method is based on a structured six-prompt reflection framework where students will be asked to explain key ideas in their own words, make connections between concepts, apply knowledge to examples or problems, identify uncertainty, and reflect on gaps or misconceptions in their understanding. In this process, the AI won’t be used to provide solutions or model answers. Instead, it will function as a questioning partner, prompting students to clarify, justify and extend their thinking, asking whether it can support deeper learning by encouraging active reflection.

Lead institution
  • Donna Johnson, Leeds Beckett University

Partner institutions
  • Victoria Bradley, Cardiff Metropolitan University
  • Sheri Scott, Nottingham Trent University
  • Claire Hawkes, Nottingham Trent University