LAWS can be programed to be humane and ethical; removes unethical human element
Kovic, 18 (Marko Kovic, Co-founder and President of ZIPAR Policy Brief, February 2018, accessed on 6-27-2022, ZIPAR Policy Brief, "The strategic paradox of autonomous weapons", file:///C:/Users/ncb12/Downloads/Policy_Brief___Autonomous_weapons.pdf) (NB)
War has always created suffering. Over the centuries, however, the morality of warfare has improved. For example, spoils of war for the victors, including raping women and pillaging, used to be common military practice. Today, there are jus in bello rules that dene morally acceptable and morally unacceptable behavior during warfare. These rules are the rules that comprise international humanitarian law, both as formal treaties and as customary law [13]. The broad strokes of humanitarian law, such as the Geneva Conventions, are almost universally accepted. However, humans make errors, and we can be irrational, emotional, sadistic, psychopathic, and so forth. Morally acceptable behavior in war is difficult to monitor and almost impossible to enforce. Autonomous weapons could make a significant positive impact in this area: Since autonomous weapons achieve their goals with the help of artificial intelligence, that artificial intelligence can be designed in such a way that moral principles of humanitarian law are at the top of the autonomous weapons’ utility function [14, 15, 16]. All the errors and biases that we humans are prone to do not exist in artificial intelligence (Unless we design the artificial intelligence in such a faulty way.). In addition, the rules of humanitarian law can be very explicitly and formally implemented in autonomous weapons, without the possibility that the autonomous weapon can override those rules. After all, artificial intelligence is simply a utility-maximizing apparatus that seeks to achieve the goals it is designed to achieve. This is a very simple point, but one that might run counter to a more intuitive understanding of morality in the context of artificial intelligence. We might feel that artificial intelligence is intrinsically «bad» or «evil» because it is only a machine, not a human. In reality, however, artificial intelligence is merely a tool that performs (some) tasks better than humans, precisely because it does not suer from our human limitations. If our goal is to wage war in a morally acceptable manner, then artificial intelligence could help us do so better. I believe that the probability for this outcome is very high, somewhere between 0.8 and 1.