Loftus v. Optum Services, Inc.

In March 2026, Sarah Loftus, a Minnesota resident, filed a class-action lawsuit against Optum Services, Inc., a subsidiary of UnitedHealth Group, after her critical post-surgical care authorization was repeatedly denied by the company's proprietary AI-based claim review platform. Loftus alleged that the algorithm overrode her treating physician's recommendations without meaningful human review, causing severe health deterioration. The complaint asserted violations of the Employee Retirement Income Security Act (ERISA) for breach of fiduciary duty, and argued that Optum's use of opaque, unaccountable software constituted a systemic pattern of denying valid claims to reduce costs, affecting thousands of plan beneficiaries nationwide. Optum defended its process as standard cost-management practice and consistent with plan terms.

Pending class certification. In a preliminary ruling, the court denied Optum's motion to dismiss, finding that Loftus plausibly alleged that the AI tool's denials without individualized physician judgment could constitute an arbitrary and capricious claims procedure. Discovery into the algorithm's design, training data, and error rate was ordered.

Whether Optum's exclusive reliance on an AI-driven prior authorization system to deny medical claims violates ERISA fiduciary duties and constitutes improper delegation of discretionary authority to an unaccountable algorithm, depriving beneficiaries of full and fair review.

The case ignited intense national debate over the use of "black box" AI in healthcare administration. It prompted proposed federal regulations requiring transparency and human override capabilities in automated claim adjudication, and spurred several states to introduce legislation mandating that AI denials be subject to independent medical review. Patient advocacy groups cited the lawsuit as a landmark challenge to corporate medicine's increasing reliance on cost-cutting algorithms at the expense of doctor-patient decision-making.