A qualitative study of 45 Shenzhen stakeholders found medical AI governance concerns centered on safety, regulatory clarity, responsibility allocation, and cross-sector cooperation. The research points to friction points that can slow adoption, especially where clinicians and local regulators need shared accountability models. The findings matter for deployments because “governance” is becoming a prerequisite for scaling clinical AI systems in routine workflows. Stakeholders also flagged the practical challenge of coordinating responsibility between developers, hospitals, and oversight bodies—an area that directly affects procurement, monitoring, and incident response.