CLEAR-AI: A clinician-led framework for responsible hospital AI governance

Artificial intelligence (AI) is increasingly embedded in hospital pathways through diagnostic platforms, triage tools, documentation assistants, scheduling algorithms, and clinical decision-support applications. These systems have the potential to improve efficiency and quality of care; however, many are adopted before clear clinical governance structures are established, creating an accountability gap. In such situations, clinicians remain responsible for decisions influenced by tools that they neither selected, validated, nor monitored. Existing frameworks, including guidance from the World Health Organization (WHO), the NIST AI Risk Management Framework, FUTURE-AI, and DECIDE-AI, emphasize trustworthy AI, transparency, human oversight, and lifecycle evaluation. However, hospitals still require practical mechanisms for assigning responsibility and accountability during implementation. This Short Communication proposes CLEAR-AI, a clinician-led governance model built around five domains: clinician-led procurement; liability and accountability mapping; evaluation before integration; audit trails and continuous monitoring; and responsible patient disclosure. CLEAR-AI is designed as a pragmatic institutional framework for the accountable adoption of AI, particularly in African and other resource-variable health systems, where digital maturity, regulatory capacity, and local validation infrastructure vary considerably. The model supports tiered, locally adaptable governance rather than silent or vendor-driven AI deployment.