Toward Trustworthy AI-Enabled Smart Healthcare: A Multi-Layer Taxonomy of Data Modalities, Model Families, Clinical Applications, and Real-World Deployment Challenges

E. A. Zanaty, Department of Computer Science, Faculty of Computers and Artificial Intelligence, Sohag University, Sohag, EgyptFollow
Asmaa H. Abd El-Rahiem, Department of Computer Science, Faculty of Computers and Artificial Intelligence, Sohag University, Sohag, EgyptFollow

ORCID

E. A. Zanaty: https://orcid.org/0009-0004-0298-8945

Asmaa H. Abd El-Rahiem: https://orcid.org/0009-0005-3947-1840

Keywords

Smart healthcare, Artificial intelligence, Clinical decision support, Wearable sensing, Explainable AI, Medical data analytics, Trustworthy AI, Digital health

Article Type

Review Article

Abstract

Artificial intelligence (AI) is increasingly reshaping healthcare through predictive analytics, medical imaging, remote patient monitoring, clinical decision support, and personalized care. Despite rapid progress, the literature on AI-enabled smart healthcare remains fragmented across data modalities, model families, clinical tasks, trustworthiness requirements, and real-world deployment challenges. This review addresses this fragmentation through a taxonomy-based synthesis of AI-enabled smart healthcare systems supported by a systematic literature search. The review proposes a multi-layer taxonomy spanning healthcare data modalities, AI model families, clinical application domains, trustworthiness attributes, and deployment readiness. The reviewed evidence covers major healthcare data sources, including electronic health records, medical imaging, wearable and physiological signals, audio data, genomic data, and multimodal data fusion. It also compares machine learning, deep learning, federated learning, explainable AI, reinforcement learning, and foundation models in terms of clinical utility, interpretability, scalability, privacy preservation, and deployment readiness. The findings indicate that AI shows promising but uneven evidence in disease prediction, image-based diagnosis, remote patient monitoring, and clinical decision support. However, integration into routine healthcare remains limited by data heterogeneity, insufficient external validation, limited model transparency, privacy and bias concerns, interoperability barriers, and regulatory uncertainty. By integrating technical, clinical, trustworthiness, and deployment perspectives, this review provides a structured framework for developing clinically validated, transparent, fair, privacy-preserving, and deployable AI-enabled smart healthcare systems.

How to Cite

Zanaty, E. A. and El-Rahiem, Asmaa H. Abd (2026) "Toward Trustworthy AI-Enabled Smart Healthcare: A Multi-Layer Taxonomy of Data Modalities, Model Families, Clinical Applications, and Real-World Deployment Challenges," Sustainable Machine Intelligence Journal: Vol. 14: Iss. 2, Article 4.
DOI: https://doi.org/10.63689/3005-3617.1088