Can we design artificial persons without being manipulative?

Abstract

If we could build artificial persons (APs) with a moral status comparable to this of a typical human being, how should we design those APs in the right way? This question has been addressed mainly in terms of designing APs devoted to being servants (AP servants) and debated in reference to their autonomy and the harm they might experience. Recently, it has been argued that even if developing AP servants would neither deprive them of autonomy nor cause any net harm, then developing such entities would still be unethical due to the manipulative attitude of their designers. I make two contributions to this discussion. First, I claim that the argument about manipulative attitude significantly shifts the perspective of the whole discussion on APs and that it refers to a much wider range of types of APs than has been acknowledged. Second, I investigate the possibilities of developing APs without a manipulative attitude. I proceed in the following manner: (1) I examine the argument about manipulativeness; (2) show the important novelty it brings to a discussion about APs; (3) analyze how the argument can be extrapolated to designing other kinds of Aps; and (4) discuss cases in which APs can be designed without manipulativeness.

Let us make man in our image-a Jewish ethical perspective on creating conscious robots

“…First, both the Golem and Adam were made using the same initial material-i.e., "the dust 2 There is a spectrum of opinions on the prospect of building conscious machines. Some hold it will happen (e.g., [110], [112], p. 39, [40], p. 37, [91], p. 375, [142]), some hold, that while it may happen, it is a long way off (e.g., [159], p. 499, [164]; p. 349, [126]; p. 146, Boden 2016 and Dennett 1997 cited in [42], p. 38); others are more conservative and hold that it is most unlikely to happen (e.g., [8], p. 164, [151], p. 467, [90], p. 444, [20], p. 210, [42], p. 66, [116], p. 2), and finally, others are emphatic that it simply won't happen (at least not computationally-see, e.g., [48,54,56], [124], p. 447, [156], p. 197, [170]). For a survey of when "High-Level Machine Intelligence" is expected to happen, see [27].…”

“…superintelligence? 14 On the one hand, proponents of conscious humanoids argue that such beings could answer very real human needs-e.g., providing the infertile a child [116], 15 the lonely a friend ( [166], p. 5), the lovesick a partner ( [94], p. 209), the world its sages [25,26,58]? On the other hand, noble though these causes are, such beings raise significant ethical concerns, for example:…”