Georgia Tech: 31 AI healthcare lawsuits reframe accountability

Two Georgia Tech researchers went through 31 US lawsuits involving AI tools in healthcare and concluded that the wrong person is usually being blamed.

In a Nature Health analysis, Gennie Mansi and Mark Riedl argue that "the medical, legal and technology communities often centre physicians' role in preventing and mitigating harms from medical artificial intelligence (AI) tools." Their read of the litigation record is that patient care depends on a "complex web of stakeholders" that includes "physicians, state health departments, health insurers and care facilities, among others," and that when an AI tool causes harm, patients have "had no option but to seek legal recourse."

The pair propose two moves. Change liability structures so that wider stakeholder set shares responsibility, and design tools that "help advocates, such as patients' lawyers, provide critical legal expertise and practically support recourse for patients." The cases are drawn from three sources: the Database of AI Litigation, the O'Neill Institute Health Litigation Tracker, and the AIAAIC incidents database, which is where the paper points readers for named systems such as nH Predict.

The abstract does not itself list the 31 lawsuits or say which stakeholder category dominates the sample.