Pacifist Military Robots

Dr Ryan Tonkens (Monash University Centre for Human Bioethics, )

ABSTRACT: In military contexts, the role of semi-autonomous robots with lethal capabilities is becoming more pronounced, and, given deep-seeded and ubiquitous assumptions driving the automated warfare enterprise, is likely to continue to expand in the future. The sophistication and degree of autonomy of the robots will likely continue to expand as well. This paper offers some pacifist reflections on the ethics of automating warfare. In it I rehearse some arguments for why all (semi-) autonomous robots ought to be pacifists—that is, why we ought to program them to be pacifists rather than warists, if we create them at all—most of which do not depend on pacifism being true: (1) human moral agents ought to be pacifists, and thus ought to program the robots they create to adhere to pacifist principles; (2) convincing robots of the moral permissibility of killing humans may be difficult to do in a way that is consistent and satisfactory; (3) there are more effective ways in which we could reduce human casualties in war, and improve the overall moral calibre of warfare, other than introducing autonomous lethal machines onto the battlefield; and (4) technology is appropriately used for serving good human ends, and killing humans (against their will) is not a good human end. I defend this view against some potential objections that have recently appeared in the literature, and strengthen it by beginning to develop two additional reasons why (semi-) autonomous robots ought not to be able to wilfully kill human beings (against their will), inside or outside the context of warfare, namely: (i) a human being murdered at the hands of a robot, whether deserving or not, is unfair, and (ii) there may be ways to design such technologies for the (laudable) purpose of effective national self-defence, without requiring that we give them lethal capabilities.