Artificial intelligence (AI), particularly deep learning (DL), has gained significant attention in ophthalmology for its potential to improve diagnostic accuracy and efficiency in detecting various eye diseases. DL has been applied to fundus photographs, optical coherence tomography (OCT), and visual fields, achieving robust performance in detecting diabetic retinopathy, retinopathy of prematurity, glaucoma, age-related macular degeneration, and other retinal diseases. DL systems can be integrated with telemedicine to screen and monitor eye diseases in primary care settings, addressing challenges in healthcare systems, especially in low-income countries. DL has shown excellent diagnostic performance in detecting diabetic retinopathy, with systems achieving high accuracy rates. For example, a DL system developed using 128,175 retinal images achieved an area under the receiver operating characteristic curve (AUC) of 0.991. Similarly, a system developed by Abramoff et al. achieved an AUC of 0.980 for detecting referable diabetic retinopathy. These systems have been validated in various studies and have received regulatory approval for clinical use. In age-related macular degeneration (AMD), DL has been used to detect and classify different stages of the disease, with systems achieving high sensitivity and specificity. DL has also been applied to OCT images to detect conditions such as choroidal neovascularisation and macular oedema, with high diagnostic accuracy. In glaucoma, DL has been used to detect the glaucoma-like disc and assess optic nerve damage. DL systems have shown promise in detecting early signs of glaucoma and predicting disease progression. However, challenges remain in terms of algorithm interpretability, data generalisability, and patient acceptance of AI-based systems. In retinopathy of prematurity (ROP), DL has been used to automate the detection of severe cases, with systems achieving high accuracy in identifying plus disease. These systems have the potential to improve screening efficiency and reduce the burden on healthcare professionals. Despite the high accuracy of DL systems, challenges remain in clinical implementation, including data variability, algorithm interpretability, and patient acceptance. Future research is needed to address these challenges and improve the clinical deployment of DL systems in ophthalmology. Overall, DL has the potential to revolutionise the practice of ophthalmology by improving diagnostic accuracy, efficiency, and accessibility.Artificial intelligence (AI), particularly deep learning (DL), has gained significant attention in ophthalmology for its potential to improve diagnostic accuracy and efficiency in detecting various eye diseases. DL has been applied to fundus photographs, optical coherence tomography (OCT), and visual fields, achieving robust performance in detecting diabetic retinopathy, retinopathy of prematurity, glaucoma, age-related macular degeneration, and other retinal diseases. DL systems can be integrated with telemedicine to screen and monitor eye diseases in primary care settings, addressing challenges in healthcare systems, especially in low-income countries. DL has shown excellent diagnostic performance in detecting diabetic retinopathy, with systems achieving high accuracy rates. For example, a DL system developed using 128,175 retinal images achieved an area under the receiver operating characteristic curve (AUC) of 0.991. Similarly, a system developed by Abramoff et al. achieved an AUC of 0.980 for detecting referable diabetic retinopathy. These systems have been validated in various studies and have received regulatory approval for clinical use. In age-related macular degeneration (AMD), DL has been used to detect and classify different stages of the disease, with systems achieving high sensitivity and specificity. DL has also been applied to OCT images to detect conditions such as choroidal neovascularisation and macular oedema, with high diagnostic accuracy. In glaucoma, DL has been used to detect the glaucoma-like disc and assess optic nerve damage. DL systems have shown promise in detecting early signs of glaucoma and predicting disease progression. However, challenges remain in terms of algorithm interpretability, data generalisability, and patient acceptance of AI-based systems. In retinopathy of prematurity (ROP), DL has been used to automate the detection of severe cases, with systems achieving high accuracy in identifying plus disease. These systems have the potential to improve screening efficiency and reduce the burden on healthcare professionals. Despite the high accuracy of DL systems, challenges remain in clinical implementation, including data variability, algorithm interpretability, and patient acceptance. Future research is needed to address these challenges and improve the clinical deployment of DL systems in ophthalmology. Overall, DL has the potential to revolutionise the practice of ophthalmology by improving diagnostic accuracy, efficiency, and accessibility.