Both the House Energy and Commerce Committee and the House Ways & Means Committee budget reconciliation bills included sections regarding artificial intelligence (AI) that LeadingAge finds worrisome. Section 43201 of the Energy and Commerce budget reconciliation bill appropriates $500 million available for 10 years to the U.S. Department of Commerce for the purposes of modernizing and securing federal information technical systems by using commercial AI, automation technology and replacing outdated systems. It is a separate clause in this section that provoked concern from numerous members of the committee due to its sweeping application and the long duration.

The concerning clause would prohibit states or other political subdivisions from enforcing existing or creating new laws or regulations regulating AI models, AI systems or automated decision systems for 10 years following enactment. Republicans on the committee argued that it makes more sense to regulate AI on the federal level than having a multitude of state policies. While this makes sense in principle, Congress has had several years to pass bipartisan recommendations governing AI and have yet to do so.

While AI shows a lot of promise, we also know that it has been misused by some Medicare Advantage plans to wrongfully deny medically necessary care and payments to providers. If the data used by the AI is not correct, the results will also be wrong. Passage of this provision would prevent states from implementing and enforcing necessary guardrails and protections as AI technologies evolve.

An attempt to remove the section was made by Democrats but voted down. There is some question whether this provision will survive a review on the Senate side under the Byrd Rule, which prohibits provisions in reconciliation bills that do not directly affect federal spending or revenues.

Section 112204 of the House Ways & Means budget reconciliation bill is more targeted appropriating $25 million to the U.S. Department of Health and Human Services(HHS) to contract with AI contractors and data scientists to implement AI for identifying and recouping improper Medicare A and B payments no later than January 1, 2027. The appropriated funds would be transferred from the Federal Hospital Insurance Trust Fund (this covers Medicare Part A) and the Federal Supplementary Medical Insurance Trust Fund (which covers Medicare Parts B and D).

While HHS would be required to report to Congress on the progress being made once the AI is implemented, it does not offer any further instructions indicating whether the recovered payments would return to the two trust funds that fund items and services for Medicare beneficiaries or be returned to cover other budget provisions.

LeadingAge is concerned that without appropriate guardrails, the AI could identify numerous payments as improper simply due to human documentation error instead of fraudulent practices. Some Medicare Advantage plans have already begun using such AI tools to kick out prior authorizations or claims where volumes of documentation have inconsistencies in the record.

We will continue to monitor the budget reconciliation package’s progress in the week’s ahead and continue to share our concerns with these and other provisions.