More than a dozen states enacted healthcare AI legislation in the first half of 2026 — most of it targeted at the place nurses have been fighting for years: prior authorization. A May 2026 review by law firm Holland & Knight documented bills in Alabama, Indiana, Utah, Washington, Maryland, and Georgia that restrict or regulate AI's role in denying care or downgrading claims. For nurses who've spent hours on hold fighting prior auth decisions, the direction of this legislation is worth knowing.
What States Are Actually Passing
The bulk of 2026's healthcare AI legislation falls into three categories:
Prior authorization restrictions. Alabama (SB 63), Indiana (HB 1271), Utah (SB 319), Washington (SB 5395), Maryland (HB 1563), and Georgia (SB 544) all restrict AI's role in the prior authorization process. The common thread: AI cannot be the sole basis for denying or downgrading a claim. A licensed professional must review the AI output and make the adverse determination independently. This directly affects case management nurses, utilization review nurses, and nurses who interact with insurance case managers on behalf of patients.
Title protection. Washington's HB 2155 (effective June 11) and Delaware's HB 191 bar AI systems from being called nurses or being licensed as nurses. Oregon passed similar legislation in 2025. The trend is accelerating: at least four additional states have similar bills in committee as of May 2026.
Behavioral health guardrails. Maine, Idaho, Nebraska, Oregon, and Tennessee restrict autonomous AI decision-making in mental health contexts, require disclosure when patients interact with AI, and in some cases mandate self-harm detection protocols. Maine's HB 2082 specifically limits AI in mental health settings to administrative functions — it cannot perform therapeutic decision-making.
Why Prior Auth Is the Real Battleground
The prior authorization fight matters most for frontline nurses because it directly translates to patient outcomes. When AI denies a medication, imaging study, or procedure — or when case managers are pressured to rubber-stamp AI denials — the nurse at the bedside is the one fielding the call from the patient who can't get their drug covered. The new state laws that require human review of AI-flagged denials put a licensed professional back in the decision chain.
How well this works in practice will depend on enforcement. A utilization review nurse who is expected to process 150 cases per day and "review" AI outputs is functionally rubber-stamping them regardless of the law's intent. What nurses and case managers should watch for is whether their employer's prior auth workflow has genuinely implemented human override checkpoints — or whether the new law becomes a compliance checkbox with no clinical substance behind it.
Where the Federal Level Stands
The White House AI policy framework released in early 2026 made no specific mention of nurses or healthcare worker implications — a gap flagged by the American Academy of Nursing and multiple nursing professional organizations. The AAN issued a comprehensive position statement in February 2026 with 13 specific recommendations on AI in healthcare, covering data privacy, algorithmic bias, workforce development, and the fundamental principle that AI must augment, not replace, nursing judgment. As of May 2026, those recommendations have not been incorporated into federal AI policy.
The American Nurses Association has been more vocal about AI guardrails than any time in the organization's history. Their 2026 position: AI tools in healthcare must be subject to nurse-led oversight, their performance must be monitored for bias, and no AI output should substitute for direct patient care assessment by a licensed nurse.
The prior auth legislation is the most practically useful development here. Nurses in case management or utilization review roles now have explicit legal backing when they push back on AI-generated denial recommendations. Know what law your state has. If your employer's workflow doesn't include genuine human review checkpoints, that's a compliance gap — and you, as the licensed professional in the chain, carry exposure if something goes wrong.