The challenges of artificial intelligence in dermatology for the immunosuppressed patient

This article explores the limitations of AI-driven dermatology tools using a patient journey as an example. Barry is a Black kidney transplant recipient with HIV whose pigmented skin cancer was initially misclassified as low-risk by an AI app. This piece highlights how current AI systems, often trained on non-diverse datasets and optimized for melanoma detection, can fail vulnerable patients: particularly those with darker skin, immunosuppression, and lesions in sensitive or hair bearing areas. The article calls for more inclusive data, ethical considerations, and human oversight to ensure equitable dermatologic care.