[PDF]

AI chatbot to provide guidance and reassurance to diabetic eye screening patients


Jack Downie

07/05/2026

Supervised by Stuart M Allen; Moderated by Mohammad Taher Pilehvar

Following diabetic eye screening, patients receive formal results letters that can be hard to interpret. For example, the overall outcome may be that no changes have been seen, some changes were observed but don’t require any further action at this time, or that the changes noticed require follow up and a further appointment with ophthalmology. The letters contain very little other information so it leaves patients stressed and anxious about whether they will lose their vision. The aim of this project is to develop a proof of concept conversational AI chatbot that provides an easy to use, accurate and verifiable interface to give patients access to further information. During the project, you will develop your knowledge and practical skills in using large language models and other AI tools, but with a strong focus on human factors and the ethical use of AI. The project will involve interaction with researchers from the Centre for Vision Services Research at Cardiff University.

https://www.cardiff.ac.uk/research/explore/research-units/centre-for-vision-services-research


Initial Plan (02/02/2026) [Zip Archive]

Final Report (07/05/2026) [Zip Archive]

Publication Form