The Ethics of the Ethics of AI

Abstract This chapter discusses several challenges for doing the ethics of artificial intelligence (AI). The challenges fall into five major categories: conceptual ambiguities within philosophy and AI scholarship; the estimation of AI risks; implementing machine ethics; epistemic issues of scientific explanation and prediction in what can be called computational data science (CDS), which includes “big data” science; and oppositional versus systemic ethics approaches. The chapter then argues that these ethical problems are not likely to yield to the “common approaches” of applied ethics. Primarily due to the transformational nature of artificial intelligence within science, engineering, and human culture, novel approaches will be needed to address the ethics of AI in the future. Moreover, serious barriers to the formalization of ethics will be needed to overcome to implement ethics in AI.