The application of artificial intelligence (AI) in dermatology has revolutionized skin cancer detection, yet significant challenges remain due to the underrepresentation of black skin in dermatological datasets. This review explores the issue of data scarcity in skin cancer detection for Skin of Color (SOC) populations, focusing on its impact on diagnostic equity and accuracy. Explainable AI (XAI) emerges as a vital approach to mitigate these biases by enhancing model transparency and accountability. Key XAI techniques, such as saliency maps, Shapley values, and gradient-weighted class activation mapping, are reviewed for their potential to improve trust and interpretability in clinical settings. Additionally, advanced methodologies like synthetic data generation and transfer learning are discussed as strategies to address data imbalance and improve model generalization across diverse skin types. By emphasizing the importance of diverse and inclusive datasets, this review highlights pathways to create AI tools that are both equitable and clinically effective, ultimately advocating for collaborative efforts between AI researchers, dermatologists, and public health stakeholders to improve skin cancer detection for all populations.