States face fragmented rules for AI health liability

A Congressional Research Service review says health-related AI liability is largely being tested in state courts under state law, even as no uniform national framework governs injuries linked to AI systems. The gap is becoming more visible as insurers use AI for prior authorization and claims review, with lawsuits alleging batch denials without meaningful physician review.

Disputes now extend beyond insurance. Courts are considering claims involving a skin assessment app under Illinois’s Biometric Information Privacy Act, facial-recognition temperature kiosks, and a chatbot used by minors for mental health interactions. Some courts have begun treating certain software as a product when it is proprietary, mass marketed, and allegedly harmful because of developer-controlled design choices, while limiting claims based on protected expression.

State lawmakers are pursuing targeted rules for AI in mental health, coverage decisions, and companion chatbots, but federal preemption remains a major constraint. Medicare preemption has blocked some state consumer protection claims against Medicare Advantage plans, while an ERISA challenge against similar claims involving private insurers did not succeed at the district court level.

The Trump administration is moving toward a minimally burdensome national AI policy and has directed the Attorney General to create an AI Litigation Task Force to challenge state laws that conflict with federal policy. The outcome will shape whether patients can seek remedies through state law and how much authority states retain over AI-driven health harms.