Senate draft proposes Medicare payment test for autonomous clinical AI

The draft would direct Medicare to test paying for autonomous clinical AI: software that can perform clinical tasks without real-time physician review.[2] The proposed test would be run through the Center for Medicare and Medicaid Innovation, which would test a payment model for autonomous AI care.[1] The proposal also seeks a federal payment pathway for AI-based services, rather than treating payment as a question separate from the legislation’s other proposed frameworks.[1] The evidence describes a proposed test; it does not specify a launch date, payment amount or final coverage decision.[1][2]

The proposal would clarify which healthcare AI tools fall under FDA oversight and which fall under ONC oversight.[1] It would give the FDA a risk-based review process for software devices and allow HHS to certify third parties to review low- and moderate-risk software.[1] HHS would also develop a coverage evaluation framework, while the proposal would set out reimbursement methodologies and create HCPCS Level II codes for AI services.[1] These provisions connect oversight and payment infrastructure, but the evidence does not detail the criteria for coverage or the specific reimbursement rates.[1]

The draft would require patient notice and preserve appeal rights.[2] It would also direct HHS to issue guidance on AI use under HIPAA and commission liability recommendations.[1] Separately, the proposal would preempt state product-liability laws for covered tools, a consequential feature for organizations assessing legal exposure.[1] Cassidy was accepting stakeholder comments on the draft through Dec. 1.[1] The evidence identifies these provisions but does not describe the notice format, appeal process or the content of future liability recommendations.[2][1]

Although the proposed payment test is for Medicare, the draft’s coverage and coding framework could eventually affect Medicaid managed-care payment for AI-enabled services.[1] That is a possible later effect, not a statement that Medicaid payment would change immediately.[1] For organizations following the proposal, the distinction between the planned Medicare test and the framework’s potential downstream reach is important when assessing budgets, implementation plans and policy engagement.[1]