Ideal theory in AI ethics

Abstract

This paper addresses the ways AI ethics research operates on an ideology of ideal theory, in the sense discussed by Mills (2005) and recently applied to AI ethics by Fazelpour & Lipton (2020). I address the structural and methodological conditions that attract AI ethics researchers to ideal theorizing, and the consequences this approach has for the quality and future of our research community. Finally, I discuss the possibilities for a nonideal future in AI ethics.

AIdeal: Sentience and Ideology

“…The AIdeal operates in the background of AI research in both industry and the academy, but also in presentations of thinking machines in popular media, tech journalism, public policy, and throughout public discussions of AI. These ideals are especially noticeable in AI ethics scholarship [57,58] , and as we have seen in this paper, they are prominent features (or perhaps, obstacles) in debates over artificial sentience and thinking machines. In this paper, we have focused on ideals that arise in the philosophical tradition around sentience and the faculties of the soul.…”

What do academics say about artificial intelligence ethics? An overview of the scholarship

“…Another way to think about AI ethics is the division of moral and perhaps legal responsibility between human and machine agents [26], and accompanying that, the potential rights and duties of machine agents [55,79]naturally, this debate raises fundamental issues around moral agency, machine sentience, and humanism [11,21,49,84,92]. Yet for many others, an ethical approach to AI requires nothing less than the critique of social, political, legal, cultural systems and institutions that maintain and reproduce relations of oppression and exploitation [12,17]. Accordingly, doing the right thing goes beyond making individual business decisions that maximize socially desirable outcomes and minimize undesirable ones; this radical critique asks AI developers and broader society to acknowledge the potential harm to the environment, vulnerable communities, and social relations, and in some cases not develop AI tools that pose unacceptable risks.…”