Philosophy for AI Ethics: Metaethics, Metaphysics, and More

Abstract

This chapter examines further philosophical issues underlying questions in AI ethics, concentrating on metaethics but also briefly examining issues in theory of knowledge, philosophy of language, and metaphysics. Many questions in AI ethics can be better understood if metaethical questions are addressed. The MIT Moral Machine experiment is analysed to introduce many of the central issues, illustrating some useful strategies for practical ethical discussion and demonstrating the necessity of considering assumptions about how moral judgements should be formed. Contrasting approaches to how we understand ethical issues are presented. We discuss ideal observer theory, the question of the universality or otherwise of moral values, the question of moral relativism, the role of reason and emotion in ethical judgement, and attempts to ground and justify moral judgement upon empirical observation or on the basis of evolution. We briefly review how questions of epistemology occur repeatedly in AI and in its ethics and present an overview of standpoint epistemology and its implications for AI. We introduce pragmatics in the philosophy of language, which is essential to understanding questions of meaning and interpretation in natural language processing. Last, the question of the nature of reality is discussed in the context of virtual reality.