INTERFACE DESIGN, ATTENTION, AND OVERRELIANCE IN CLINICAL XAI

Background: Clinical explainable artificial intelligence (XAI) is commonly evaluated as explanation content: text, probability, saliency map, feature ranking, counterfactual, or analogous case. In practice, an explanation is encountered through an interface that controls when advice appears, where attention is directed, how much effort verification requires, and whether acceptance or rejection is documented. Materials and Methods: This narrative review synthesized DOI - inde Xed literature on clinician-centred XAI design,trust in AI-based clinical decision support systems, eye tracking, explanation type, cognitive forcing, false confirmation, advice-taking, overreliance, and reporting guidance for clinical AI.