AI Insurance Denials Under Fire: Why States Are Demanding Human Review

The Bipartisan Revolt Against Algorithmic Denials

A rare point of bipartisan agreement is reshaping American health insurance: legislators from red and blue states alike are telling insurers they cannot use artificial intelligence as the sole basis for denying medical claims. As of April 2026, more than ten states have either enacted laws or advanced legislation mandating that a licensed physician must review any AI-driven coverage denial before it reaches a patient.

The movement was triggered by high-profile controversies. UnitedHealth Group’s nH Predict algorithm and Cigna’s automated denial systems drew national attention when investigations revealed that AI tools were rejecting claims at industrial scale with minimal or no physician oversight. Patients with serious conditions found their coverage denied by algorithms that never reviewed their medical records.

What the Laws Actually Require

The enacted statutes share a common architecture: AI can assist in claims processing, but a qualified human clinician must make the final determination on any denial.

California led the charge with SB 1120, the Physicians Make Decisions Act, effective January 1, 2025. The law prohibits any denial, delay, or modification of care based on medical necessity unless reviewed and decided by a licensed physician with expertise in the relevant clinical area.

Arizona followed with a law effective July 1, 2026, requiring a licensed Arizona physician to personally review and sign off on any AI-based denial of claims or prior authorizations involving medical necessity.

Nebraska’s LB 77 prohibits AI output from being the sole basis for evaluating medical necessity to deny, delay, or modify healthcare services. Maryland’s HB 820 adds an audit dimension, requiring that AI tools used for utilization review incorporate individual medical history and remain open for state inspection.

Utah, Connecticut, and Texas have enacted comparable restrictions, with Texas requiring written disclosure to patients when AI is used in connection with healthcare services, effective January 1, 2026.

The 2026 Legislative Wave

The momentum is accelerating. Georgia’s SB 444 passed the Senate and requires human clinical review before insurers deny doctor-ordered care. Minnesota is advancing a bill specifically targeting AI denials of prior authorization requests. Kansas, Pennsylvania, and several other states have introduced similar measures in their 2026 sessions.

The numbers tell a striking story: 47 states introduced more than 250 healthcare AI bills during the 2025 legislative session alone, with 33 signed into law across 21 states.

Federal vs. State Tension

What makes this wave particularly significant is the emerging tension with the federal government. The White House has signaled disagreement with the state-level approach, favoring industry self-regulation and lighter-touch federal guidelines. This creates a potential preemption battle that could define the future of AI governance in healthcare.

Insurance industry groups argue that AI improves efficiency, reduces processing times, and can catch fraud more effectively than manual review. They warn that mandating human review for every denial could slow claims processing and increase administrative costs. However, patient advocacy organizations counter that speed cannot come at the cost of accuracy when lives are at stake.

The Compliance Challenge for Insurers

For health insurers operating across multiple states, the emerging patchwork of laws creates significant compliance complexity. Each state defines “human review” slightly differently — California requires a physician with relevant clinical expertise, Arizona mandates a licensed state physician, and Maryland focuses on audit transparency. Insurers must now map their AI claims workflows against at least ten distinct regulatory frameworks, with more coming online quarterly.

The practical impact is forcing insurers to redesign their AI-assisted claims pipelines. Rather than eliminating AI entirely, most are building hybrid workflows where algorithms flag potential denials but licensed clinicians make the final call. This “AI-assisted, human-decided” model is emerging as the de facto compliance standard.

Sources & Further Reading

• Landmark Law Prohibits Health Insurance Companies from Using AI to Deny Healthcare Coverage — Senator Josh Becker
• States Pursue Legislation Limiting AI’s Growing Role in Payer Prior Authorization Denials — Dark Daily
• Red and Blue States Want to Regulate AI in Insurance. The White House Disagrees — Governing
• 47 States Introduced Healthcare AI Bills in 2025 — Becker’s Hospital Review
• Georgia Bill Would Ban AI-Only Insurance Denials — Hoodline
• Briefing Book 2026: AI Use in Health Insurance — KLRD