Related Experiment Video
Updated: Aug 12, 2026
03:14
Augmenting Large Language Models via Vector Embeddings to Improve Domain-Specific Responsiveness
Published on: December 6, 2024
Evaluation of artificial language-adjusted readability on cataract surgery queries
Michael X Lin1, Azam S Husain1, Libby Wei1
1University of Maryland Department of Ophthalmology and Visual Sciences, Baltimore, Maryland.
Digital Journal of Ophthalmology : DJO
|August 11, 2026
Summary
Artificial intelligence (AI) models can simplify cataract surgery information for better understanding. These AI tools maintain accuracy while improving the readability of complex medical queries.
Area of Science:
- Ophthalmology
- Artificial Intelligence
- Medical Informatics
Background:
- Cataract surgery is a common procedure with complex information.
- Patient comprehension of medical information is crucial for informed decision-making.
- Artificial intelligence (AI) tools are increasingly used to generate health-related content.
Purpose of the Study:
- To evaluate the accuracy and readability of AI-generated answers to cataract surgery questions.
- To determine if AI models can significantly enhance the readability of medical information.
Main Methods:
- Three AI models (Google Gemini Advanced, ChatGPT 4.0, Microsoft Copilot Pro) answered 25 cataract surgery questions.
- Answers were generated at a standard and a 6th-grade reading level.
- Readability was assessed using validated formulas and expert ophthalmologist grading.
- Accuracy and readability were compared between original and simplified versions and across AI models.
Main Results:
- AI models successfully lowered the reading level of cataract surgery information.
- Microsoft Copilot Pro and ChatGPT 4.0 showed greater improvements in readability compared to Google Gemini Advanced.
- Ophthalmologist grading confirmed no loss of accuracy when simplifying the text.
Conclusions:
- AI models can effectively simplify complex cataract surgery queries for improved patient understanding.
- Readability of AI-generated medical information can be enhanced without compromising accuracy.