What is it Like to Encounter an Autonomous Artificial Agent?

Abstract

In his seminal paper “What is it like to be a bat?” published in 1974 Thomas Nagel argues in favor of the irreducibility of the first-person perspective. His main argument is that being a bat, a spider, an ape, or a human being implies that those particular living beings experience something and that this something exists in his own right. Following Nagel, it is simply a fact that this or that living being has this or that experience; it is a fact that cannot be denied by arguing that such experiences can be described just by using physical terms which do not provide for a space for an entity called mind or private respectively subjective experiences named qualia. Roughly spoken, in his paper Thomas Nagel rejects the idea of eliminative physicalism. Although Nagel’s major aim is to argue in favor of the irreducibility of the first-person perspective he also makes another point that might be fruitful in the discussion concerning whether autonomous artificial agents could be conceived as morally responsible actors. This argument or, better to say, this intuition is presented in merely one sentence in Nagel’s essay. After some introductory paragraphs he says: “Even without the benefit of philosophical reflection, anyone who has spent some time in an enclosed space with an excited bat knows what it is to encounter a fundamentally alien form of life.” (Nagel 1974: 438) In what follows firstly it shall be shortly analyzed what the above cited sentence might mean or, better to say, what conclusions could be drawn from it concerning the question about the relationship of human beings on the one and autonomous artificial agents on the other side. In a second step some empirical studies concerning the behavior and reactions of human beings who were confronted with different kinds of autonomous artificial agents shall be presented to show that particularly the appearance of such agents determines the way human beings react towards them. Finally, some tentative conclusions shall be drawn with regard to the design of autonomous artificial agents. It will be argued that designers, manufacturers and service providers might be tempted to use designs that divert the assignment of moral responsibility to the autonomous artificial agents instead of to these designers, manufacturers and service providers themselves.