Artificial intelligence-generated informed patient consent in various ophthalmological procedures: A comparative study of correctness, completeness, readability, and real-word application between Deepseek and Chatgpt 4o
View abstract on PubMed
Summary
This summary is machine-generated.AI chatbots generated informed consent forms for eye surgery. While Deepseek offered better readability, 40% of consents from both ChatGPT and Deepseek were unsuitable for the Indian context.
Area Of Science
- Ophthalmology
- Medical Informatics
- Artificial Intelligence
Background
- Informed consent is crucial for ethical medical practice.
- The use of artificial intelligence (AI) in healthcare is rapidly expanding.
- Evaluating AI-generated medical documentation is essential for patient safety and legal compliance.
Purpose Of The Study
- To assess the quality of AI-generated informed consent forms for ophthalmological procedures.
- To evaluate correctness, completeness, readability, and real-world applicability of AI-generated consents in the Indian context.
Main Methods
- A cross-sectional observational study involving ophthalmology faculty.
- AI chatbots (ChatGPT 4o, Deepseek) generated informed consents for common ophthalmic procedures.
- Generated consents were evaluated for correctness, completeness, readability, and applicability in India.
Main Results
- Deepseek provided more readable consents than ChatGPT, with higher readability scores.
- Both chatbots required multiple attempts to generate suitable responses.
- A significant 40% of AI-generated consents were deemed unfit for use in the Indian scenario.
Conclusions
- AI chatbots like Deepseek offer more readable informed consents compared to ChatGPT.
- Currently, both AI chatbots fail to consistently generate appropriate informed consent forms for the Indian healthcare setting.
- Further refinement of AI models is needed to ensure the clinical and legal suitability of generated medical documentation.

