Hey Watson – Can I Sue You for Malpractice? Examining the Liability of Artificial Intelligence in Medicine

Abstract

IBM’s Watson achieved celebrity on the quiz show Jeopardy! but it is now weighing in on matters of life and death as an expert system in the healthcare setting. Increasing automation in areas such as oncological research and diagnostics promises signi cant reductions in morbidity and mortality, but several barriers stand in the way of such technology’s widespread implementation and acceptance. As Watson achieves greater accuracy and is increasingly relied upon by doctors, one key question will become: how should liability be assigned where medical errors can be traced to an “arti cial” intelligence (AI)? Using philosophical and legal concepts, this article classi es Watson’s current capabilities and examines how humans make “good” decisions in order to better characterize Watson’s role and potential faults. The authors conclude that Watson warrants a unique legal status akin to personhood and is analogous to a medical resident, in that both require “oversight” by an attending physician. We therefore argue that liability for wrongful diagnoses by medical AI should attach on a medical malpractice basis rather than through a products liability or vicarious liability scheme, and can thus be addressed contractually between the AI provider and medical institutions rather than constituting an open question for legislatures.