What happened
On 29 October 2026 Senate HELP Committee Chairman Bill Cassidy released a legislative discussion draft (not yet introduced or voted) called the PATIENT AI Act, aimed at expanding access to AI in health care. It would direct the CMS Innovation Center to run a five-year payment model for AI-based applications capable of performing a clinical function autonomously, restate that Medicare cannot reject a service 'solely because' AI works without simultaneous physician review, create an FDA risk-based review process for software devices (allowing certified third-party review of low/moderate-risk software), and call for HIPAA privacy guidance for AI use. It would require informed beneficiary notice and clinical accountability/escalation arrangements. Public comments are due 1 December 2026.
Why it matters
This is the first detailed legislative proposal to create a Medicare reimbursement model for autonomous clinical AI, which would materially change the economics and regulatory treatment of clinical AI agents. It also sets out an FDA risk-based software pathway and pushes on the unresolved liability question for autonomous AI, though it sends liability to further study rather than resolving it. It signals a shift toward federal payment and oversight for AI that performs clinical tasks without a human review at each step.
Action needed
Stakeholders in AI-enabled health care should review the draft and submit comments by 1 December 2026; device makers and clinical-AI vendors should track the FDA risk-based software review and Medicare payment-model provisions.