Hospitals push for rule clarifying human review when insurers use AI in prior‑authorization denials

Cheyenne — A bipartisan group of providers and insurers told the Labor, Health & Social Services Committee on June 24 that prior authorization denials must include clinician review and that proposed legislation should clarify how insurer use of artificial intelligence (AI) fits into existing law. “What we bring before you today is a small clarification to that law,” Josh Hannes of the Wyoming Hospital Association told the committee, asking the panel to add language that would require a licensed clinician to “meaningfully review” any insurer decision premised on an AI output before an adverse determination is issued. Hospitals raised several cases from other states in which employees alleged AI was being used to triage or drive denials; Hannes and others asked the Legislature to make a technical change to Wyoming Statute 26‑55‑104 so the statute explicitly requires a documented clinical review when an insurer’s denial traces back to an algorithmic recommendation. Blue Cross Blue Shield of Wyoming’s Kelsey Prestesater told the panel the company has already committed not to use AI to make prior‑authorization denials in Wyoming and affirmed the company’s support for transparent rules. “Blue Cross Blue Shield of Wyoming already has committed to not utilizing artificial intelligence in the prior authorization process for adverse determinations,” Prestesater said. Several carriers and trade representatives including Cigna, Mountain Health Co‑op and CVS/Caremark — and the Wyoming Department of Insurance — said the current statute already requires a clinician to make adverse determinations, and that the department enforces the statute through complaints and market conduct. The Department of Insurance noted it has not received complaints alleging insurers in the state are using AI to deny prior authorization but said the NAIC is studying AI and that state departments are coordinating on model language. Committee members asked practical questions about enforcement and whether a clinician who merely “rubber‑stamps” an AI recommendation would meet statutory standards. Bruce Spencer, an attorney representing Mountain Health Co‑op, cautioned against inserting unclear language that would inadvertently “muddle” an otherwise clear statutory requirement that an adverse determination be made by a physician or appropriate health care provider. Hospitals countered that the clarification sought was purposefully narrow: it would not ban insurer use of AI as a triage tool but would require documentation that a licensed and clinically knowledgeable provider reviewed any adverse decision that originated with AI and that the reviewer considered clinical context. Why it matters: Committee discussion reflected a national policy debate about algorithmic decision‑making in health care and the adequacy of clinician oversight in insurer utilization management. Insurers emphasized existing legal obligations and voluntary corporate commitments; hospitals said clarification will give clinicians and patients greater certainty about who made a denial and why. The committee asked stakeholders to work together on language and flagged the issue for future consideration; members suggested holding the draft for the general session so stakeholders and regulators can align on model language developed by NAIC and federal guidance.