We Are Not Pontius Pilate: Acknowledging Ethics and Policy

Abstract

A new AI system is being developed to optimize vaccination strategies based on the structure and shape of a community's social contact network. The technology is minimally constrained and not bound by preconceived notions or human biases. With this come novel outside the box strategies; however, the system is only capable of optimizing what it is instructed to optimize, and does not consider any ethical or political concerns. With the growing concern for systematic discrimination as a result of artificial intelligence, we acknowledge a number of relevant issues that may arise as a consequence of our new technology and categorize them into three classes. We also introduce four normative ethical approaches that are used as a framework for decision-making. Despite the focus on vaccination strategies, our goal is to improve the discussions surrounding public concern and trust over artificial intelligence and demonstrate that artificial intelligence practitioners are addressing these concerns.