This article discusses a study that used non-mydriatic fundus cameras and AI software to detect diabetic retinopathy in patients. The study found that the software was able to accurately detect pathological findings in fundus images, potentially aiding in early detection and treatment of the disease.