Supportive AI in Dermatology Evidence, Limitations, and Governance for Latin America

Dermatology is one of the most visible areas of artificial intelligence in health, because many initial decisions depend on images, morphological patterns, and timely triage. Published evidence shows sustained progress in skin-lesion classification, differential-diagnosis support, teledermatology, and human-model collaboration, but it does not justify promises of replacing medical judgment or deployments without governance. Recent reviews indicate that the largest volume of evidence concentrates on skin cancer and dermoscopy, with less maturity for inflammatory diseases, longitudinal follow-up, and populations with skin of color. The focused search identified no prospective clinical validations of dermatology AI in Latin America; it did find regional experience in teledermatology and dermoscopy training that informs feasibility, not AI effectiveness. For clinical teams in Latin America, AI can help organize visual and textual information, prioritize cases, support clinical education, and improve decision traceability when integrated with human review. It cannot yet promise reliable autonomous diagnosis across every skin type, capture device, or care setting. Implementation gaps include representative data, local validation, consent, cybersecurity, regulation, workflow integration, and professional accountability. This community whitepaper summarizes the current state of evidence for four tasks: lesion and skin-cancer classification, broad dermatologic diagnostic support, teledermatology and triage, and data and implementation governance. This narrative synthesis examines supported use cases, unsupported claims, and priorities for local research.