Artificial intelligence and medical retina training: A scoping review of educational opportunities, emerging risks, and curricular responses

Autonomous artificial intelligence (AI) systems for retinal image interpretation are being deployed in routine clinical practice, fundamentally altering the training environment of retinal specialists. We map available evidence on how AI integration affects medical retina specialist training, addressing educational opportunities, developmental risks, and curricular responses. Following the PRISMA extension for scoping reviews and Joanna Briggs Institute methodology, we searched PubMed/MEDLINE, Embase, Web of Science, the Cochrane Library, and gray literature from major ophthalmological and medical education organizations from inception to March 2026. Six thematic domains emerged: deployment context, artificial intelligence as educational tool, explainability and pedagogy, risks to trainee development, large language models and assessment validity, and institutional responses. AI creates genuine opportunities for personalized case allocation, synthetic dataset generation, explainable visual feedback, and knowledge scaffolding in medical retina training, while simultaneously posing documented risks including deskilling, never-skilling, automation bias, and disruption of established assessment frameworks. The professional identity formation of trainees in environments where AI routinely matches or exceeds first-year resident performance remains an underexplored concern. Institutional and accreditation responses lag substantially behind clinical deployment, ophthalmology-specific competency frameworks validated for residency training are largely absent, and targeted research and coordinated curriculum reform are urgently needed.