Draft PATIENT AI Act would clarify FDA or ONC oversight and add AI payment pathway

Sen. Bill Cassidy, MD, who chairs the Senate Health, Education, Labor and Pensions Committee, released a discussion draft of the PATIENT AI Act dated Oct. 29, according to Becker's Hospital Review. The draft would settle a question many health systems face today: which federal body oversees a given AI tool. AI applications that do not qualify as FDA-regulated devices would fall under ONC's certified health IT program. FDA would gain a risk-based review process for software-based devices, including authority to accept premarket software changes so updated tools reach patients faster, and HHS would certify third-party bodies to review low- to moderate-risk software. The draft also restates federal preemption of state product liability laws for these devices.

The reimbursement provisions matter most to finance leaders. HHS would build a coverage evaluation framework for AI-based services, propose payment methodologies and establish HCPCS Level II codes for them. The Center for Medicare and Medicaid Innovation would test a payment model for AI applications that deliver care autonomously. HHS would also issue HIPAA compliance guidance for AI use and commission a nonprofit to recommend approaches to AI liability. Stakeholder feedback is open through Dec. 1 at patientai@help.senate.gov.

The jurisdiction question is live outside Congress as well. STAT reported on Oct. 30 that FDA and CMS officials are working to head off conflict as more states look to test AI in clinical settings. States license clinicians and regulate the practice of medicine, while AI that performs clinical tasks such as writing prescriptions without human supervision would likely fall under FDA device jurisdiction.

What the evidence shows is limited. This is a discussion draft, not law, and no outcomes exist. Codes, payment rates and the scope of the autonomous care model are proposals that could change or never take effect. What it does signal is the direction of federal thinking on oversight, liability and payment, and a window to shape it.

The move for the Chief AI Officer, CFO and compliance lead is a joint inventory before Dec. 1. Sort each AI tool in use or in procurement into three groups: likely FDA-regulated device, likely ONC certified health IT, and neither. Flag which services could plausibly qualify for a coverage framework or code, and which operate with minimal human supervision. That map supports an informed comment letter, clarifies vendor contract terms on regulatory status and liability, and positions revenue cycle teams to act if a payment pathway becomes real. For clinicians and staff, clearer oversight could mean fewer stalled pilots and more confidence that the tools they are asked to use have a defined accountability path.

You are an AI governance and reimbursement advisor to the executive team of [type of organization, such as a health system or multi-site practice].

Context: Sen. Bill Cassidy, MD, who chairs the Senate HELP Committee, released a discussion draft of the PATIENT AI Act dated Oct. 29, according to Becker's Hospital Review. AI applications that do not qualify as FDA-regulated devices would fall under ONC's certified health IT program. FDA would gain a risk-based review process for software-based devices, and HHS would certify third-party bodies to review low- to moderate-risk software. HHS would build a coverage evaluation framework for AI-based services, propose payment methodologies and establish HCPCS Level II codes. The Center for Medicare and Medicaid Innovation would test a payment model for AI applications that deliver care autonomously. Feedback is open through Dec. 1 at patientai@help.senate.gov. Separately, STAT reported on Oct. 30 that FDA and CMS officials are working to head off conflict as more states look to test AI in clinical settings. This is a discussion draft, not law. No outcomes exist, and codes, payment rates and model scope could change or never take effect.

Task: Using my list, sort each AI tool in use or in procurement into three groups: likely FDA-regulated device, likely ONC certified health IT, or neither. For each tool, flag whether it could plausibly qualify for a coverage framework or code, and whether it operates with minimal human supervision.

Output: a one-page memo for the executive team with these sections: Inventory map (a table), Payment pathway candidates, Minimal supervision flags, Vendor contract questions on regulatory status and liability, Points for a comment letter before Dec. 1, and Items to verify.

My inputs: [your organization], [your list of AI tools in use and in procurement, with vendor and function], [the named owners for Chief AI Officer, CFO and compliance lead], [your current vendor contract terms on regulatory status].

Guardrails: Use only the facts above and the inputs I fill in. Cite Becker's Hospital Review and STAT by name. Flag every classification that is uncertain or needs legal or regulatory verification. Use no patient data. Nothing here is medical or legal advice. Label the memo as a draft for the team to check.