Autonomous weapons systems and the application of IHL
Host item entries:
Collegium, No. 41, Automne 2011, p. 71-77
The author discusses the following question: Can we take the rules of IHL and apply them to robots in the battlefield? In the case of the principle of distinction, three characteristics help us distinguish between an enemy and a non-combatant: the physical element, such as the uniform and whether they are carrying a weapon. Then there are the behaviour characteristics, which is the type of movement. Then the geographical characteristics: where are they in relation to the military target? Although a technical challenge, it is feasable to train a machine to learn these three characteristics. However, the question is whether it will be done or not, because it will cost a lot of money and it will be very complicated. It might also create a situation where robot makers may have more information than they want to. It could be that they do not want to know so much about their weapon systems on the battlefield, because then, they could be liable. Proportionality can be broken down to two principles: Distinction and the balance between military advantage and collateral damage. If we want to teach a robot, we will have to develop a formula or a system for them to mimic our decision-making process. These are things that we have been working on for fifty years and we are not there yet.