Conjectures of the Possibility of Recasting Hume’s Guillotine for Hybrid-Designed Artificial Moral Agents

This article explores the prospects of re-appropriating Hume's Guillotine to argue against the suggestion of developing artificial moral agents via a hybrid approach.Hume's Guillotine states that any type of evaluative claim may never be exclusively drawn from factual statements.Hybrid routes of constructing artificial moral agents, meanwhile, are those that fuse two kinds of design approaches, namely: (1) top-down programming, which entails encoding artifacts with a set of moral precepts, and (2) bottom-up strategies that are characterized by learning or developmental techniques.In this essay, it is argued that there seems to be some reasonable grounds for believing that artifacts made by means of hybrid methods are vulnerable to Hume's Guillotine, especially in terms of their ability to conjure moral judgments.To preliminarily lay the groundwork for this claim, certain objections posed against top-down and bottom-up strategies are revisited in the hopes of further recasting them (i.e., to show that the moral reasoning faculty of systems built through hybrid processes are also suspect).