A new study from the University of Cambridge found that OpenAI's GPT-4 model performs close to the level of expert ophthalmologists in analyzing eye conditions and suggesting treatments.