A recent decision in Estate of Lokken v. UnitedHealth Group, Inc. highlights emerging litigation risks associated with insurers’ use of artificial intelligence tools to guide coverage decisions. The case involves allegations that UnitedHealth Group relied on an AI-driven predictive model developed by its subsidiary naviHealth to determine the appropriate duration of post-acute medical care for insureds.
The court’s recent ruling addressed a discovery dispute and allowed plaintiffs to obtain substantial information about the development, operation, and oversight of the AI system. The permitted discovery includes materials relating to how the model was designed and trained, what data it relied upon, how its outputs were used in claims decisions, and whether human reviewers exercised meaningful independent judgment. The ruling underscores that courts may allow deep scrutiny of an insurer’s algorithmic decision-making processes when plaintiffs allege that automated tools are influencing coverage determinations.
Although the case arises in the health insurance context, the underlying theories could have broader implications across the insurance industry. For property insurers in particular, the decision is a reminder that automated estimating tools, damage-detection technologies, underwriting risk models, and fraud analytics systems may become focal points in future litigation. As AI adoption continues to expand across the insurance sector, carriers should ensure that these tools are deployed with appropriate human oversight, clear documentation, and governance structures capable of withstanding potential legal scrutiny.
This post was originally published through Horst Krekstein & Runyon’s Property in 60 Seconds Newsletter. If you would like to receive future copies of that newsletter, please use this link to sign up.