Explainable Artificial Intelligence: Building Trust in the Sleep Clinic

Artificial intelligence (AI) is increasingly applied in sleep medicine, where complex multimodal data create opportunities for automation and clinical decision support. However, many high-performing AI models remain black-boxes, limiting transparency, trust, and safe clinical adoption. Explainable AI (XAI) addresses this gap by providing insight into model behavior, helping identify bias and assess unreliable predictions. XAI also supports emerging regulatory expectations for transparency and human oversight. This review summarizes key XAI principles, methods, limitations, regulatory considerations, and clinical applications in sleep medicine, including sleep staging and sleep apnea diagnostics, emphasizing AI as a tool to augment, rather than replace, clinical expertise.