Three More States Bar Insurers from Using AI as the Sole Reason to Deny Coverage

By Hector Herrera | April 14, 2026 | Health

Medical disclaimer: This article covers health insurance policy and regulation. It is informational and does not constitute medical or legal advice.

Nebraska, Maryland, and Maine enacted laws last week prohibiting health insurers from using AI as the sole basis for denying or modifying coverage claims, according to the Troutman Privacy Blog. The three states join Indiana, Utah, and Washington, which passed similar laws earlier in 2026 — bringing the total to six states with explicit AI health insurance restrictions in a single year.

What the Laws Do

The laws share a common structure: an insurer cannot issue a denial or modification of a claim solely on the basis of an AI system's determination. A human with clinical expertise must review any AI recommendation before a coverage decision takes effect.

This is a narrower restriction than it might appear. The laws do not prohibit AI in insurance claims processing. They do not require AI recommendations to be disclosed to patients. They do not regulate AI-assisted pricing or underwriting. What they prohibit specifically is fully automated adverse decisions — a denial that happens because an AI flagged it and no human reviewed it.

The specific practice being targeted is what the industry calls "automated prior authorization": AI systems that review claims against coverage criteria and issue instant decisions. When those systems deny claims without human review, patients can face delays in care or outright refusals with no human ever having evaluated their case.

Context

AI in health insurance claims is not hypothetical. Multiple major insurers — including UnitedHealthcare, Cigna, and Humana — have faced litigation and regulatory scrutiny over AI claims processing systems that generated denial rates that plaintiffs alleged were designed to minimize payouts rather than accurately assess medical necessity.

A 2024 Senate investigation found that UnitedHealthcare's AI prior authorization tool, nH Predict, had a denial override rate above 90% — meaning that when patients or doctors appealed AI denials, more than 9 in 10 were reversed. A 90%+ override rate on appeal is a strong signal that the original denial was wrong. When wrong denials are generated at scale by an automated system, patients who don't appeal — and many don't, because the appeals process is difficult — receive less care than they are entitled to.

The state laws are a direct response to that documented pattern.

The Federal Collision

The cluster of state legislation is arriving at the same time the White House is pushing for a federal AI framework intended to preempt state AI laws — meaning federal rules would supersede state rules, and states could not impose more stringent requirements.

The stated justification for federal preemption is regulatory consistency: AI companies and insurers operating nationally find it difficult to comply with a patchwork of 50 different state frameworks. The argument has merit in theory.

In practice, the preemption push arrives at a moment when the White House's posture on AI regulation is deregulatory. A federal framework designed to preempt state patient protection laws — written by an administration skeptical of AI regulation — is likely to provide weaker protections than the state laws it would displace.

Over 600 AI bills have been introduced across U.S. state legislatures in 2026. Most will not pass. But the cluster of health insurance AI laws — six states in one year — reflects a specific area where state legislators see concrete constituent harm and are acting on it.

What Patients Need to Know

If you receive a denial for a medical procedure, treatment, or prescription from your health insurer, you have the right to appeal. In the six states that have passed these laws, you now also have the right to have a qualified human review any AI-generated denial before it stands.

Even outside those six states, most insurance policies have appeal processes that require human review at some stage. Use them. The documented 90%+ reversal rate on AI denials in some systems suggests appeals are more likely to succeed than many patients assume.

Key questions to ask when you receive a denial:

• Was this decision made by an AI system or reviewed by a human clinician?
• What specific coverage criteria did my claim fail?
• What documentation would be needed to support an appeal?

What to Watch

The federal preemption effort will be the defining battle. If Congress passes federal AI legislation that preempts state health insurance AI laws, the six existing state laws — and any others passed in 2026 — could be voided. Watch for the American Hospital Association, patient advocacy groups, and state attorneys general to challenge any preemption that weakens current state protections.

Also watch for more state action. With six laws passed in four months and 600+ bills in play, the question is not whether more states will act on health insurance AI — it is how many will do so before federal preemption resolves the jurisdictional question.

Hector Herrera covers AI in healthcare and health policy for NexChron.