The Role of Large Language Models in Ophthalmology: Opportunities, Risks, and the Critical Need for Safeguards

1

Large language models (LLMs) like GPT-4 show potential in ophthalmology for documentation and patient communication but pose significant risks.
2

LLMs do not integrate real-time clinical data, leading to confident yet potentially inaccurate responses in patient assessments.
3

Hallucination remains a critical issue, with LLMs generating plausible but factually incorrect content that can harm patients.
4

Biases in training datasets may result in LLM outputs being less accurate for under-represented demographic groups in ophthalmology.
5

Current governance for LLMs in clinical settings is lacking, necessitating mandatory auditing and oversight before deployment.

1