Turki Alelyani
Artificial Intelligence (AI)-powered autonomous systems are increasingly entering healthcare, yet concerns about their reliability, safety, and responsible use present significant barriers to adoption. Building on prior conceptual work, this study introduces a refined and empirically validated framework designed to support the safe and responsible integration of AI in clinical and regulatory contexts. The original framework was developed from semi-structured interviews with 15 experts across clinical, technical, ethical, and regulatory domains, and was subsequently validated through a structured process involving 10 newly recruited participants. Validation combined quantitative ratings and qualitative feedback, yielding consistently high scores for relevance, clarity, and usability, alongside strong endorsement of practical utility. The resulting framework consists of ten dimensions spanning technical, ethical, and operational categories, and is aligned with international standards such as ISO 21448 and the NIST AI Risk Management Framework. By addressing critical issues including data quality, explainability, fairness, and human-AI collaboration, the framework moves beyond abstract principles to provide actionable guidance. It offers clinicians, developers, regulators, and procurement bodies a structured tool to evaluate, monitor, and guide the responsible adoption of autonomous AI systems in healthcare ecosystems.