A large language model artificial intelligence (AI) system can match or even outperform human ophthalmologists in the diagnosis and treatment of patients with glaucoma and retinal disease. The research from New York Eye and Ear Infirmary of Mount Sinai, published in JAMA Ophthalmology, indicate that advanced AI tools could play an important role in providing decision-making support to ophthalmologists.

The study matched the knowledge of 15 ophthalmic specialists against the capabilities of the latest generation AI system GPT-4 from OpenAI. It matched or outperformed human specialists in accuracy and completeness of medical advice and assessments. AI demonstrated superior performance in response to glaucoma questions and case-management advice, but in retina, humans matched AI in accuracy but exceeded it in completeness.