Talking Eyes: Large Language Models for Patient Education in Ophthalmology

The purpose of this study is to investigate the potential of large language models (LLMs), including ChatGPT and Google Gemini, to improve health literacy in ophthalmology. By analyzing their ability to provide personalized responses to health-related inquiries, this research aims to assess the effectiveness of these AI tools in addressing ophthalmic conditions.

This narrative review analyzed 31 studies evaluating the performance of publicly available LLMs in patient education across various ophthalmic topics. Key parameters assessed included accuracy, accessibility, and adaptability to diverse populations.

LLMs produced accurate responses on eye health, simplifying medical information and adapting the readability and language of its outputs. Challenges included occasional misinformation, misunderstood contexts, and Western-centric biases. LLM performance improved when specific strategies were used, such as prompt engineering and revisions of patient education materials by ophthalmologists.

LLMs' accessibility harbours the potential to improve ophthalmic health literacy. To maximize their utility, collaborative development with patients and ophthalmologists, legislative oversight, and continuous updates on ophthalmic data are essential.