AMA publishes governance framework for agentic AI taking on clinical functions
The American Medical Association's non-binding framework holds that AI's clinical authority should be earned through evidence, and sets four questions on function, evidence and risk, oversight and accountability.
The American Medical Association (AMA) has published a governance framework for agentic AI systems that assume clinical functions. According to Robinson+Cole, it was released on 13 October. Its starting principle is that clinical authority should be “earned through evidence and maintained through demonstrated performance”, with authority and oversight able to change as evidence, performance, risk or the technology evolve.
The framework is built around four connected questions: what clinical function the AI performs; what evidence shows it can do so and what happens if it fails; what it should be permitted to do and under which conditions; and who is accountable for risks and outcomes. According to Robinson+Cole, accountability is to be assigned by each participant’s authority, control, risk knowledge and ability to prevent harm.
The guidance is not binding. The AMA says the questions will shape its work with physicians and others on liability, licensure and regulation.
Why it matters
The framework offers health organisations a template for an AI use-case inventory that records each agent's function, intended use, patient population and setting. AI governance teams can map its evidence-and-risk question to validation and monitoring records scaled to the consequences of failure, and its accountability question to named owners with real control over each agent. Treating agent permissions as conditional on ongoing performance evidence implies periodic reassessment rather than one-off approval.
Status: Published. Non-binding framework for health care providers
Sources
Was this useful?
Thanks. What was the problem?
Thank you — this helps us choose what to cover. If you spotted an error, please tell us what it is.
Prepared with AI assistance and reviewed by the editor before publication. Spotted an error? Tell us — see our editorial policy and methodology.
Governance areas
Who should care
- AI governance lead
- Chief medical information officer
- Legal & compliance
- Risk managers