Summary
Organizes governance of clinical AI agents around function, evidence and risk, authority and oversight, and accountability. Calls for evidence proportionate to consequence and autonomy, responsibility aligned with control and ability to prevent harm, and continued monitoring, learning, and reassessment.
Healthcare Implications
Provides clinicians, health systems, developers, payers, and policymakers with a voluntary framework for defining and revisiting an AI agent's clinical role and safeguards. Its discussion of licensure means authorization for a defined AI capacity, not a license to practice medicine; it does not change legal scope-of-practice rules.