Artificial intelligence (AI) holds potential in dermatology education, but existing literature demonstrates high risk of bias, has small sample sizes, and is characterized by limited methodological rigour, according to a review by Singapore researchers. Furthermore, evidence to determine whether AI-based tool improve learning outcomes compared with standard educational approaches is lacking.
“While current initial studies on AI hold promise, this scoping review underscores the need for more robust studies with standardized evaluation frameworks, prioritizing ethical principles such as fairness, explainability, and accountability for safe integration in training,” wrote Chun Hian John Yap from the Ministry of Health Holdings, and Ziying Vanessa Lim from the National Skin Centre.
Yap and Lim did a systematic search of PubMed, Embase, Web of Science, Scopus, and PsycINFO up to July 2025 for studies on AI in dermatology education. They synthesized data narratively, considering the types of AI interventions and evaluated outcomes using the CUC-FATE framework, and assessed study quality using ROBINS-I and a COSMIN-informed checklist.
A total of 827 records were identified, of which 61 were assessed for eligibility. After screening, only six full-length studies and one conference abstract, mostly from 2023 to 2025 met the inclusion criteria. [Ann Acad Med Singap 2026;55:370-382]
Eligible studies examined AI-generated clinical images, Large Language Model-generated vignettes, intelligent tutoring systems, and clinical decision support tools. In content validation studies, ratings for accuracy, clarity, and educational utility were generally favourable. Furthermore, intervention studies reported potential benefits for learning performance and diagnostic accuracy.
Reviewed studies commonly assessed usability and credibility of AI, while cost, accountability, fairness, transparency, and explainability were hardly explored.
“Most studies were preliminary, small-scale and heterogeneous in design, and current evidence is insufficient in supporting conclusions regarding its effect on improving learner performance and clinical competence,” said Yap and Lim.
Medical education
These findings support emerging evidence from the broader medical education literature, with an increasing number of AI-related publications over the years. [Med Clin Pract 2024;7:100422]
A scoping review of AI medical image generation across health professions education also stated that generative AI should be treated as an experimental adjunct that needs thorough human review. [J Med Syst 2026;50:19]
Another scoping review of AI-based approaches in radiology education reported possible applications, including personalized curriculum generation and diagnostic support tools, as well as automated evaluation systems. [Curr Probl Diagn Radiol 2025;54:332-338]
“While similar themes were observed in the present review, the evidence base identified for dermatology was considerably smaller, with fewer studies evaluating educational outcomes and learner performance,” Yap and Lim said.
“Together, these reviews suggest that while dermatology shares characteristics with more visual specialties—making it potentially well-suited for AI-enhanced learning—robust evidence demonstrating educational effectiveness is lacking, making further research with rigorous validation necessary to identify the most effective AI tools in education,” they added. [Surg Innov 2024;31:415-23]
The current review did not include other potential intervention, such as personalized learning platforms, integrating AI with virtual reality-based training for skin cancer screening, and AI-assisted learning for knowledge and milestone assessment, which could be included in future implementation studies to assess educational outcomes, according to Yap and Lim.