US state legislatures are accelerating efforts to limit the role of artificial intelligence in health insurance determinations, with four additional states passing significant new legislation in 2026.
Georgia Senate Bill 444, effective January 1, 2027, expressly authorises AI systems to automate tasks and reduce administrative burdens within health insurance processes.
However, the Georgia law prohibits AI from issuing an adverse determination until a natural person qualified as a private review agent or utilisation review entity conducts a review with clinical peer participation.
Critically, an AI system in Georgia must not supersede the judgment of a clinical peer, a restriction that mirrors the approach previously adopted by Pennsylvania.
Iowa House File 2635, effective July 1, 2026, permits utilisation review organisations to use AI for the initial review of prior authorisation requests under certain conditions.
However, the Iowa law is clear that AI may not be the sole basis for a decision to deny, delay, or downgrade a prior authorisation request based on medical necessity.
Denials or downgrades in Iowa must be made by a qualified reviewer or clinical peer, and utilisation review organisations must provide written explanations for any such decisions, including the reasoning and criteria used.
Utah’s Senate Bill 319, enacted on March 19, 2026 and taking effect January 1, 2027, amends the state’s health care preauthorisation statute to require that individuals making adverse determinations based on clinical or medical necessity exercise independent medical judgment.
The bill prohibits reliance solely on “recommendations from any other source,” a provision that directly captures AI-generated recommendations, and requires insurers to disclose any use of AI to the Utah Insurance Department.
Utah insurers will also be required to post conspicuously on their websites all preauthorisation requirements, including a clear explanation of how AI is used in the review process.
Washington’s Senate Bill 5395, effective June 11, 2026, takes one of the most comprehensive approaches yet, governing AI use by health carriers, health care benefit managers, and public employee health plans.
Under the Washington law, only a licensed physician or licensed health professional acting within their scope of practice may deny a prior authorisation request based on medical necessity.
Human reviewers in Washington must evaluate the requesting provider’s recommendation, the enrollee’s medical or clinical history, and the enrollee’s individual clinical circumstances before issuing a denial.
Washington also requires that AI operating in prior authorisation processes comply with applicable law, be used in a fair and nondiscriminatory manner, and be subject to audit by the state’s insurance commissioner.
Carriers operating in Washington must additionally report to the insurance commissioner the percentage of total prior authorisation denials that were aided by AI, adding a layer of public accountability absent from the other states’ legislation.
On the legislative pipeline, Pennsylvania House Bill 1925 remains with the legislature after being laid on the table as of May 5, 2026, while Oklahoma House Bill 3675 also remains pending.
Bills in both New Hampshire and Louisiana have not progressed, with Louisiana Senate Bill 246 withdrawn by the Louisiana Senate on May 28, 2026, and New Hampshire House Bill 1406 also dying without passage.