Health Insurer AI Use Creates Friction
Health insurers are increasingly adopting AI to reduce expenses and improve productivity, but the move is sparking tension with healthcare providers. Providers have expressed concerns about a lack of trust and the potential for opaque, "black box" decisions in areas like claims processing and patient care authorization. The conflict highlights the growing need for transparency and explainability in insurance AI models.
- A class-action lawsuit alleges UnitedHealth Group used an AI model called nH Predict, developed by its subsidiary naviHealth, to systematically deny care to elderly patients in Medicare Advantage plans. The lawsuit claims the company knew the model had a 90% error rate, evidenced by the high number of denials that were overturned on appeal. - Cigna faces a similar class-action lawsuit for allegedly using an algorithm known as PxDx to automatically deny large batches of claims without individual physician review. According to the complaint, this system allowed Cigna doctors to reject over 300,000 claims in a two-month period, spending an average of just 1.2 seconds on each case. - Humana has also been sued for using the same naviHealth nH Predict algorithm to systematically deny medically necessary rehabilitation care for its Medicare Advantage members, overriding physicians' recommendations. - In response to the growing use of AI, the National Association of Insurance Commissioners (NAIC) adopted a model bulletin in late 2023 to guide insurers. Now adopted by at least 24 states, it requires insurers to create a written program for AI governance, risk management, and bias testing to ensure compliance with unfair trade practice laws. - Several states are now enacting "human-in-the-loop" laws. For example, an Arizona law effective mid-2026 will prohibit insurers from denying coverage based solely on AI output, requiring a licensed physician to review and approve any adverse determination. - Lawsuits argue that insurers have a financial incentive to use these systems because very few members appeal denials. One analysis found that in 2021, members appealed only 0.2% of denied claims. - According to a survey from the American Medical Association, 61% of physicians are concerned that the use of AI by health plans is leading to an increase in prior authorization denials. More than a quarter of physicians reported that a prior authorization delay has led to a serious adverse event for a patient in their care. - Despite the controversy, AI adoption is widespread, with one survey finding that 84% of health insurers use AI or machine learning in some capacity. Insurers report that nearly 92% of them have AI governance principles in place that align with the NAIC's guidelines.
