Drawing a Non-Consequentialist Line: Augmenting v. Replacing the Professions with Computer Systems

Abstract

This paper draws a line between those occupations that may eventually lend themselves to wholesale AI replacement on consequentialist grounds, in contrast to those that should not. This requires demarcating the professions from other types of expert service providers. Of those expert services whose safe delivery is in the public interest, healthcare, legal/financial services and education stand out because they give rise to a very particular type of vulnerability, one that potentially threatens the moral equality of those seeking those services (the particular case of those shaping the architecture conditioning our virtual interactions will also be discussed). This paper delineates the ethical demands stemming from this vulnerability, which is contrasted to the kind of vulnerability that characterises all lay-expert relationships given their inherent knowledge asymmetry. This contrast turns out to matter both conceptually and normatively: the legitimacy of – less and less hypothetical – wholesale replacement of some professionals by computer systems indeed hinges upon this contrast between different kinds of vulnerability. When our life is in the hands of an expert, our vulnerability may in some cases be lessened by wholesale AI replacement. Yet when, aside from this “primary” vulnerability, a professional can affect our less tangible ability to be the author of our own lives (in a way that commands respect both from ourselves and others), replacing such professionals with computer systems would imperil our commitment to moral equality – no less.

Keywords: artificial intelligence, augmentation, expertise, habit, professions, habituation, ethics, expert systems, vulnerability, automation bias