Unveiling Artificial Intelligence’s Diagnostic Power and Challenges in Dermatology for Skin of Color: A Review
January 2026 | Volume 25 | Issue 1 | e22 | Copyright © January 2026
Published online December 27, 2025
Ugwumsinachi D. Nwaubani MSc MIBa, Adaora Ntukogu BAb, Nyla Griffith BAc, Aishat Awe BSd, Sharon A. Glick MD FAADe, Antonella Tosti MD FAADf
aState University of New York Health Sciences University College of Medicine, Brooklyn, NY
bMedical College of Georgia, Augusta, GA
cMorehouse School of Medicine, Atlanta, GA
dMeharry Medical College, Nashville, TN
eDepartment of Dermatology, State University of New York Downstate School of Medicine, Brooklyn, New York
Department of Dermatology, Maimonides Medical Center, Brooklyn, NY
fUniversity of Miami Miller School of Medicine; Dr. Phillip Frost Department of Dermatology and Cutaneous Surgery; Miami, FL
INTRODUCTION
Artificial intelligence (AI) has the ability to revolutionize the diagnostic outcomes and accessibility of dermatology.¹ This study scrutinizes the application of AI in dermatology, specifically assessing its effectiveness in diagnosing skin conditions in skin of color (SOC), while synthesizing the emerging trends of dermatologic AI and SOC, thus serving as a resource to address the bias toward lighter skin tones-a fundamental shortcoming of AI- which is vastly understudied.² The goal is to help guide researchers and developers on ways to create more inclusive tools for equitable diagnostic use, highlighting limitations and providing insight into ways to mitigate bias and make change toward equitable and reliable AI dermatologic outcomes.
