무슨 일이 있었나
On 13 October 2026 the AMA's Center for Digital Health and AI released a governance framework for agentic AI systems assuming clinical functions. It is grounded in the principle that clinical authority should be earned through evidence and maintained through demonstrated performance, and focuses on four questions: function (what clinical role), evidence + risk (validation and safeguards), authority + oversight (what the system may be permitted to do and under what conditions), and accountability (who is responsible for risks and outcomes). It will guide AMA work on liability, licensure and regulation.
왜 중요한가
As agentic AI takes on consequential clinical tasks, this provides a physician-professional-association governance/assurance framework addressing permission limits, oversight and accountability — relevant inputs to AI-security/governance controls and to clinical liability and regulatory approaches for healthcare AI.
필요한 조치
Healthcare organisations and clinical-AI developers should map agentic-AI governance, permission scoping and accountability questions to this framework and to emerging clinical-AI regulatory expectations.