State AI Health Liability Laws Face Federal Preemption Push

Fragmented state laws leave AI health harms in legal limbo

Most litigation over artificial intelligence-related injuries in healthcare unfolds in state courts under state law, yet no uniform national framework exists to govern these disputes. A Congressional Research Service report examining state-level AI liability frameworks in health reveals a fundamental gap between existing legal structures and the novel harms emerging from algorithmic decision-making.

The Trump administration has signaled its intent to challenge this patchwork through executive action, directing the Attorney General to establish an AI Litigation Task Force aimed at overturning state AI laws deemed inconsistent with federal policy. The move sets up a direct collision between state experimentation in consumer protection and federal preemption efforts.

Why it matters

Healthcare AI systems are making consequential decisions about coverage and care without clear legal accountability. The absence of uniform standards leaves patients uncertain about their rights when algorithms deny claims or provide harmful advice, while insurers and developers face unpredictable liability across jurisdictions. The federal preemption push could eliminate state protections before Congress establishes baseline safeguards.

Insurers face claims over algorithmic denials

Health insurers have deployed AI tools to automate prior authorization and claims review processes. Plaintiffs in multiple lawsuits, including Estate of Barrows v. Humana and Kisting-Leung v. Cigna, allege that insurers used algorithms to deny batches of claims without meaningful physician review. These cases represent an emerging category of litigation over coverage determinations made through AI systems.

Beyond insurance, consumer-facing health applications are drawing legal scrutiny. A skin assessment app called Skin360 faced challenges under Illinois's Biometric Information Privacy Act. Temperature screening kiosks with facial recognition capabilities prompted similar litigation. In Garcia v. Character Technologies, Inc., plaintiffs challenged a general-purpose chatbot designed to interact with minors on mental health topics, claiming unsafe design.

Courts begin treating AI as a product

Historically, courts treated software as a service or declined to impose liability based on First Amendment concerns. Recent rulings show a shift. Some courts now hold that software qualifies as a product when it is proprietary in design, mass-marketed in the stream of commerce, and causes alleged harm based on design elements the developer controls.

In the Character Technologies case, the court found the AI chatbot was a product regarding design defects such as lack of age verification, while dismissing claims based on the chatbot's expressions. The same court determined Google could face liability as a component part manufacturer because its large language model was integrated into the chatbot.

Federal preemption complicates state enforcement

Plaintiffs have invoked state consumer protection laws—Unfair and Deceptive Acts and Practices statutes that every state has enacted—to challenge AI-driven health decisions. However, federal preemption presents obstacles. Courts have found that Medicare's express preemption provision blocks state UDAP claims against Medicare Advantage plans. The Employee Retirement Income Security Act, which includes a preemption provision with a savings clause, did not preempt similar claims against private health insurers, according to one district court.

State legislatures are enacting targeted measures addressing AI in mental health, coverage determinations, and companion chatbots. Arizona H.B. 2175 represents state-level legislation on AI in coverage decisions. At the federal level, S. 5117, the Senior Chatbot Protection Act of 2026, would regulate general-use chatbots for senior users while preserving state protective laws.

What comes next

The Trump administration's executive order directs federal efforts toward a "minimally burdensome national policy framework for AI" intended to sustain U.S. global AI dominance, while excluding child safety protections from its preemption scope. Most health-related AI litigation remains in early stages with no merit findings yet, leaving unclear whether state law can provide adequate remedies to injured consumers.

The coming months will determine whether Congress establishes baseline protections, whether courts uphold state laws against federal challenge, or whether competing standards persist. The resolution will shape whether injured patients have meaningful recourse and whether states retain authority to protect residents from AI-driven health harms.

These details were first reported by AI Watch through the Legis1 platform.