From Black Boxes to Dialogue: Explainable AI Agents in Dermatological Diagnosis and Patient Consultation

Dermatology is a field that relies heavily on visual pattern recognition, clinical examination, and patient history to diagnose and manage skin, hair, and nail disorders. With advances in Artificial Intelligence (AI), deep learning models trained on vast image datasets have achieved impressive accuracy in tasks such as skin lesion classification, acne grading, and eczema detection. Despite these successes, the opaque nature of many AI algorithms—often referred to as “black-box” models—has raised concerns about transparency, accountability, and usability in clinical environments. Explainable Artificial Intelligence (XAI) has therefore emerged as a critical paradigm to bridge the gap between algorithmic performance and clinical trust, enabling interpretability and facilitating collaboration between clinicians and AI systems. In parallel, intelligent conversational agents are being integrated into teledermatology and digital health platforms, serving as interactive assistants for both patients and practitioners. Yet, few studies have combined these two lines of innovation within dermatology, leaving a gap at the intersection of explainability, human–AI interaction, and clinical reasoning.This Research Topic aims to bring together interdisciplinary research exploring the design, development, and evaluation of explainable AI agents for dermatological consultation and clinical decision support. The goal is to enhance diagnostic transparency, support clinician decision-making, and improve patient understanding by merging XAI techniques with interactive AI system design. Central questions include: How can explainability be embedded within multimodal AI pipelines that integrate imaging and clinical text? What forms of explanation are most effective for dermatologists and patients? How can conversational interfaces enhance user trust and accessibility? The objective is to create systems that not only predict but also justify, facilitate dialogue, and promote safe, equitable use of AI in dermatology.To gather further insights into the evolving landscape of explainable dermatological AI, we welcome articles addressing, but not limited to, the following themes:o Development of XAI-enabled deep learning models for dermatological image analysiso Conversational and interactive AI agents for clinical decision support and teledermatologyo Multimodal AI systems combining image, text, and patient-reported datao Techniques for model interpretability, visualization, and uncertainty quantification in skin diagnosiso Bias detection, fairness, and generalization across diverse skin tones and populationso Human–AI interaction and trust calibration in dermatological consultationo Ethical, legal, and privacy considerations in AI-driven healthcareo Design of transparent decision pipelines and clinical explanation frameworkso Integration of explainable AI into electronic health record (EHR) or mobile consultation workflowso Evaluation frameworks for usability and clinical adoption of XAI-based systems Topic editors: Arvind Mukundan, Romeo Giuliano, RIYA KARMAKAR Source: https://www.frontiersin.org/research-topics/81905/from-black-boxes-to-dialogue-explainable-ai-agents-in-dermatological-diagnosis-and-patient-consultation