Review of Trustworthy AI in Digital Health: Robustness and Explainability

A recent article on arXiv (2608.02238) highlights the urgent requirement for reliable AI in the realm of digital health, emphasizing the importance of robustness and explainability. The authors contend that fostering trust in AI systems is vital for their ethical and safe application in critical fields such as healthcare. They outline essential aspects of trustworthy AI, which include robustness, explainability, fairness, accountability, and privacy, all of which must be considered from the initial problem identification to data gathering, model implementation, and human engagement. Although previous works have touched on various elements of trustworthy AI, a concentrated examination of robustness and explainability specific to healthcare has been absent. This review addresses that need by structuring recent progress into an easily navigable framework, underscoring both technical and practical aspects. It aims to assist researchers and practitioners in creating dependable and understandable AI solutions for digital health. The article has been released on arXiv as a new submission and serves as a valuable resource for those involved in AI within healthcare, synthesizing existing knowledge and pinpointing future research directions.