The Machine Question: Critical Perspectives on AI, Robots, and Ethics

Notes

Notes Introduction 1. This concept of “pure agency,” although excluded from further consideration by Floridi and Sanders, will turn out to be operationalized by functionalist approaches to designing artificial autonomous agents (AAAs). This development will be explicitly analyzed in the consideration of machine moral agency. 1 Moral Agency 1. The fact that it is not explicitly identified as such could be taken as evidence of the extent to which the instrumental definition has become so widely accepted and taken for granted as to be virtually transparent. 2. In her later work, Johnson has increasingly recognized the complexity of agency in situations involving advanced computer systems. “When computer systems behave,” she writes in the essay “Computer Systems: Moral Entities but Not Moral Agents,” “there is a triad of intentionality at work, the intentionality of the computer system designer, the intentionality of the system, and the intentionality of the user” (Johnson 2006, 202). Although this statement appears to complicate the human-centric perspective of computer ethics and allow for a more distributed model of moral agency, Johnson still insists on the privileged status and position of the human subject: “Note also that while human beings can act with or without artifacts, computer systems cannot act without human designers and users. Even when their proximate behavior is independent, computer systems act with humans in the sense that they have been designed by humans to behave in certain ways and humans have set them in particular places, at particular times, to perform particular tasks for users” (ibid.). 3. Although Kant, unlike his predecessors, Descartes and Leibniz in particular, does not give serious consideration to the possibility of the machina ratiocinatrix, he does, in the Anthropology (2006), entertain the possibility of otherworldly nonhuman rational beings, that is, extraterrestrials or space aliens. See David Clark’s “Kant’s Aliens” (2001) for a critical investigation of this material. 218 Notes 4. The role of human responsibility in this matter would then be more complicated. It would not be a question of whether human designers and operators use the object in a way that is responsible; rather, as Bechtel (1985, 297) describes it, “the programmer will bear responsibility for preparing these systems to take responsibility.” Or as Stahl (2006, 212) concludes, rephrasing the question of computer responsibility by referring it elsewhere, “can (or should) man assume the responsibility for holding computers (quasi-)responsible?” 5. There is something of an ongoing debate concerning this issue between John Cottingham and Tom Regan. Cottingham, one of Descartes’s translators and Anglophone interpreters, argues, in direct opposition to Regan and Singer’s Animal Rights and Human Obligations (1976), that animal rights philosophers have unfortunately employed a misconstrued version of Cartesian philosophy. “The standard view,” Cottingham (1978, 551) writes, “has been reiterated in a recent collection on animal rights [Regan and Singer 1976], which casts Descartes as the villain of the piece for his alleged view that animals merely behave ‘as if they fell pain when they are, say, kicked or stabbed.’ . . . But if we look at what Descartes actually says about animals it is by no means clear that he holds the monstrous view which all the commentators attribute to him.” In response to this, Regan (1983, 4) partially agrees: “Cottingham , then, is correct to note that, as in his letter to More, Descartes does not deny that animals have sensations; but he is incorrect in thinking, as he evidently does, that Descartes thinks that animals are conscious.” 6. This is already evident in Descartes’s text, in which the terms “soul” and “mind” are used interchangeably. In fact, the Latin version of the Meditations distinguishes between “mind” and “body” whereas the French version of the same text uses the terms “soul” and “body” (Descartes 1988, 110). 7. Floridi and Sanders provide a more detailed account of “the method of abstraction ” in their paper “The Method of Abstraction” (2003). 8. In taking mathematics as a model for revising and introducing conceptual rigor into an area of philosophy that has lacked such precision, Floridi and Sanders (2004) deploy one of the defining gestures of modern philosophy from Descartes to Kant but obviously extending as far back as Plato and into contemporary analytic thought. 9. There is a certain intellectual attraction to repositioning Immanuel Kant as an engineer. For instance, the First Critique is, as Kant had described it, nothing less than an attempt to reengineer philosophy in order to make it function more effectively and efficiently. “In fact...