Grant: Project Grant

Advocates argue that artificial intelligence (AI) technologies will transform health care for the better by, e.g., reducing high levels of medical error that occur within hospitals. However, there are also many concerns. For example, can a surgeon obtain "informed" consent from a patient if the surgeon does not understand how AI makes its decision? Who should be held to account, and for what, when injury arises from the use of AI? How can we prevent "algorithmic bias," which is when AI products discriminate because the data they are built upon are under-representative of, e.g., women, Indigenous Peoples or Black Canadians? And will patient privacy be jeopardized as AI tools process huge quantities of personal health information and share that data with others? We do not yet know whether Canada's complex web of uncoordinated laws and regulations will adequately protect patients and other users against potential problems with AI in health care. Our multidisciplinary research team will rigorously assess 10-12 AI case studies in different care settings, representative of the potential risks of AI in health care. We will use a case study method to scrutinize the sufficiency of legal governance of current and emerging AI technologies in the Canadian system and in selected foreign jurisdictions. This will enable us to clearly identify where there are gaps in existing legal protections and if law reform is required, whether that be changes to existing laws or ethical codes, or entirely new laws. To generate the best advice on law reform options, we will work closely not only with AI innovators and health care providers that are part of our research team, but also with key decision-makers, like Health Canada, and with patient representatives. Our goal is for Canadian patients and society to truly benefit from the revolutionary potential of AI in health. This will require optimal regulation of heath care, which is as important as the health innovation itself.