FDA and CMS Signal Joint Push to Work With States on AI Health Tech Oversight

For the millions of Americans who now have an AI tool somewhere in their medical care, whether reading a scan, flagging a billing code, or suggesting a diagnosis, the question of who is actually watching these systems has never been simple. Two federal agencies are now suggesting they want to watch together, and they want states at the table too.

The Food and Drug Administration and the Centers for Medicare and Medicaid Services are signaling a new effort to collaborate with state governments on oversight of artificial intelligence in health technology. The move comes as AI tools proliferate across clinical and administrative settings faster than any single regulator can track alone. It also comes at a politically charged moment. The announcement landed on Election Day, a timing that underscores just how much health policy, and the infrastructure behind it, can shift depending on who holds power in Washington and in state capitols.

Think of the current regulatory landscape like a house with multiple locks but no master key. The FDA has traditionally overseen AI as it shows up in medical devices, the digital equivalent of inspecting a new piece of equipment before it reaches a hospital floor. CMS, meanwhile, controls how Medicare and Medicaid pay for that equipment and the services tied to it, essentially deciding which tools get a financial green light to scale. States, for their part, license providers, regulate insurance practices, and increasingly write their own AI-specific rules. When those three layers do not talk to each other, gaps open up, and patients are the ones who fall through them.

Why Coordination Is Becoming Unavoidable

AI in health care has outgrown the narrow lane regulators originally built for it. A tool that started as a radiology aid for spotting tumors can now influence prior authorization decisions, staffing models, and even which patients get flagged as high risk for readmission. Each of those applications touches a different part of the regulatory system, and each part has historically moved at its own pace.

That fragmentation has real consequences. A hospital system deploying a new AI triage tool might clear FDA review but still face a patchwork of conflicting state disclosure requirements, or find that CMS reimbursement rules have not caught up to how the tool is actually used. Patients, meanwhile, often have no idea which layer of government, if any, is checking whether the tool works as promised for people who look like them, live where they live, or carry the same chronic conditions.

States have not waited around for Washington to settle these questions. Several have already passed their own AI transparency and accountability laws, particularly around mental health chatbots and algorithmic decision-making in insurance claims. That state-level activity creates both an opportunity and a headache for federal regulators. It means there is already a body of practical experience to draw from, but it also means any federal collaboration effort has to reconcile rules that were not designed with each other in mind.

This matters beyond the wonky mechanics of agency memos. When AI tools influence a cancer diagnosis or a decision about whether a Medicaid patient qualifies for home care, the stakes are not abstract. They are a family waiting on a result, a clinician trusting a recommendation they may not fully understand, a budget office deciding whether a cost-saving algorithm is worth the risk of getting it wrong. Coordinated oversight, done well, could mean fewer gaps for unsafe or biased tools to slip through unnoticed. Done poorly, it could mean duplicated paperwork without any real improvement in safety, a classic case of more process but not necessarily more protection.

The broader health policy environment makes this collaboration feel almost inevitable. The same week this effort emerged, STAT reported separately on how the 2024 electoral map tracks closely with a widening mortality gap between red and blue states, and on how the basic economics of American health care remain, in plain terms, broken. Health technology regulation does not exist in a vacuum. It is layered on top of a system already struggling with uneven outcomes and unsustainable costs, which raises the bar for what any new oversight framework needs to accomplish.

There is also a workforce and innovation angle worth weighing. Developers building AI health tools have long complained that inconsistent state rules make it hard to scale a product nationally, the health tech equivalent of needing a different driver's license for every state you cross. A more unified federal-state approach could ease that burden, but only if the collaboration produces clearer rules rather than another layer of ambiguity. Industry will be watching closely to see whether this becomes a genuine simplification or just an additional checkpoint.

What Comes Next

The details of how FDA, CMS, and individual states will actually structure this collaboration remain thin, and the full scope of the plan is tied up behind reporting that has not yet been made broadly available. What is clear is the direction of travel: federal regulators are acknowledging that AI in health care cannot be governed effectively from Washington alone, and states are not willing to simply wait for federal rules to catch up.

For patients and clinicians, the practical test will be whether this coordination translates into tools that are safer, more transparent, and more consistently reviewed, no matter which state a patient happens to live in. For now, the signal is that regulators see the current patchwork as unsustainable. Whether the fix arrives in time to keep pace with how quickly these tools are already shaping real medical decisions is the question worth watching closely in the months ahead.