Proxy Prudence: Rethinking Models of Responsibility for Semi­‐Autonomous Robots

20 Pages Posted: 28 May 2014 Last revised: 21 Aug 2014

Date Written: March 12, 2014

Abstract

As robots become more autonomous — capable of acting in complex ways, independent of direct human interaction — their actions will challenge traditional notions of responsibility. How, for example, do we sort out responsibility when a self-driving car swerves this way or that in a situation where all possible outcomes lead to harm? This paper explores the question of responsibility from both philosophical and legal perspectives, by examining the relationship between designers, semi-autonomous robots and users. Borrowing concepts from the philosophy of technology, bioethics and law, I argue that in certain use contexts we can reasonably describe a robot as acting as a moral proxy on behalf of a person. In those cases I argue it is important to instantiate the proxy relationship in a morally justifiable way. I examine two questions that are helpful in determining how to appropriately instantiate proxy relationships with semi-autonomous robots, and that we can also ask when attempting to sort out responsibility: 1) On whose behalf was the robot acting?; and 2) On whose behalf ought the robot to have been acting?

Focusing on proxy relationships allows us to shift our focus away from a strictly causal model of responsibility and focus also on a proxy model informed by an ethical analysis of the nature of the designer-artefact-user relationship. By doing so I argue that we gain some traction on problems of responsibility with semi-autonomous robots. I examine two cases to demonstrate how a shift towards a proxy model of responsibility, and away from a strictly causal model of responsibility helps to manage risks and provides a more accurate accounting of responsibility in some use contexts. I offer some suggestions how we might decide whom a robot ought legitimately to be acting on behalf of, while offering some thoughts on what legal and ethical implications my argument carries for designers and users.

Keywords: robot ethics, design ethics, biomedical ethics, roboethics, robot policy, moral proxy, engineering ethics, autonomous cars, driverless cars

JEL Classification: K13, L92

Suggested Citation

Millar, Jason, Proxy Prudence: Rethinking Models of Responsibility for Semi­‐Autonomous Robots (March 12, 2014). Available at SSRN: https://ssrn.com/abstract=2442273 or http://dx.doi.org/10.2139/ssrn.2442273

Jason Millar (Contact Author)

University of Ottawa ( email )

57 Louis Pasteur
Ottawa, ON K1N 6N5
Canada

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
179
Abstract Views
823
rank
196,169
PlumX Metrics