AI is a cornerstone of digital transformation and has gained increasing importance in the healthcare sector,especially in diagnosis, prediction, and clinical decision-making. These developments are characterized by technological evolution, domains of implementation, and accountability considerations (ethics, fairness, and explainability). In this regard, this study seeks to bridge the gap through a structured narrative review of AI in healthcare. The final papers of relevance have been sourced from a defined set of search strings and inclusion criteria, and screened using a transparent selection process. The selected body of literature was evaluated according to a conceptual framework with three key elements: Foundations, Applications, and Accountability. There have been significant advances in technology for prediction and diagnosis, while the use of accountability tools has been inconsistent. For example, explainability and anti-bias techniques have not been fully integrated into the processes but addressed on an ad hoc basis. This review thus provides a roadmap of the discipline with respect to technology and accountability aspects to guide the development of clinical AI responsibly.
Pradhan et al. (Sun,) studied this question.