The convergence of genomics, information sciences and Electronic Health Records is a main driver of precision medicine (PM). A common feature of definitions of PM is the use of big data for prediction or classification of groups of people for treatment, diagnosis, or prognosis. Achieving the vision of PM, with its vast data requirements, relies on the success of artificial intelligence (AI), especially for the predictive analytics needed. There are three distinctive features of AI in PM that motivate the need for ELSI inquiry: 1) The process of developing AI for PM is technically and organizationally complex, requiring acquisition of data from multiple sources and multiple types of expertise, including software engineers and computer scientists who are not familiar with the regulatory and ethical frameworks that guide medicine. 2) Sources of systematic bias in AI models for PM have been identified, but responsibility for preventing discriminatory decision-making and action on the basis of biased AI is not established. 3) Developers and users of AI for PM have divergent interests and needs. Since one of the stated goals of PM is to reduce healthcare costs, such interests can conflict with patient interests and negative effects are difficult to mitigate because of the opacity and lack of clear lines of responsibility for uses of AI. Panelists will draw from recent empirical studies of AI development and normative analysis to propose new models for data access and the process of AI development to enhance the production of ethical AI for precision medicine.
Panelists: Pamela Sankar, PhD; David Magnus, PhD; Ariadne Nichol, BA; Mildred Cho, PhD