US Health Insurers Under Fire for AI Role in Claims Decisions
US health insurers are facing increased pressure and legal challenges concerning their use of artificial intelligence (AI) in making claims decisions. Class action lawsuits and government investigations suggest that these technologies, while offering potential benefits, may be leading to inappropriate denials of medical coverage, raising serious ethical and practical concerns.
One of the central cases involves Gene Lokken, a 91-year-old man who fractured his leg and ankle. Despite his doctor’s recommendations, Lokken’s insurer, UnitedHealth, refused to continue paying for his care in a nursing facility, even though he was still recovering and in significant pain. The family had to pay over $12,000 per month until his death a year later.
Lokken’s case is one of several at the heart of class action lawsuits, which claim that insurers are using AI to reject claims, even when they are supported by a physician’s assessment. The core allegation is that the AI systems are overriding the clinical judgment of medical professionals.
Potential for Algorithmic Errors
While AI shows promise in healthcare for applications such as diagnosis and drug research, critics point to significant risks. Some argue that AI’s opaque decision-making processes can lead to errors and unjust denials.
The primary defendant in similar lawsuits is UnitedHealth, the largest insurance company in the US, with approximately 53 million policyholders through its UnitedHealthcare division. Although other major insurers like Humana are also affected, the legal and political focus is intense on UnitedHealth.
UnitedHealth uses a computer model called nH Predict, developed by its subsidiary NaviHealth. Litigants claim that this AI tool inaccurately determines the length of post-care reimbursement and limits the ability of staff to override its decisions. UnitedHealth denies these claims, and a response is scheduled for the end of March.
Regulatory Responses and Concerns
Recognizing the potential for misuse, the former US administration introduced voluntary operating agreements with insurers in 2023 and issued an executive order on AI standards in 2024.
The Centers for Medicare & Medicaid Services (CMS), which oversees the US federal Medicare health insurance system, issued guidance requiring Medicare Advantage plans to make coverage determinations based on individual circumstances rather than solely on an algorithm’s output. Assessments must be reviewed by a healthcare professional.
In October, a US Senate investigations committee raised concerns about rising rejection rates for “pre-authorization” reimbursement under the Medicare Advantage program. “Many of the issues that most frustrate patients and providers remain cloaked in uncertainty. This is particularly true of insurers’ use of automation and predictive technologies,” the committee concluded.
Insurance Company Responses and Broader Criticism
UnitedHealth has disputed the claims made against it, stating that coverage decisions are made by medical directors, not AI, and follow CMS Medicare coverage criteria. The company maintains that it pays 98% of valid claims.
However, other experts are raising additional concerns regarding AI’s impact on healthcare reimbursement. A report by the University of Chicago’s NORC research center warned of the risks of AI exacerbating biases, misaligned incentives, and the use of technologies outside their original design.
The National Association of Insurance Commissioners (NAIC) issued a Model Bulletin reminding insurers to comply with all insurance laws and regulations. A NORC submission in November noted that while some states have regulated AI in health insurance, most efforts have not kept pace, calling for transparency and human oversight.
The Trump administration has revoked the AI order, replacing it with a call for a new action plan. The legal landscape is changing, and has led to uncertainty among organizations using AI. The battle over the use of AI in healthcare is far from over, with legal and political battles continuing to shape standards and practices.
