To start with, States should assess, through legal reviews of new weapons and new methods of warfare, whether the level of human involvement in the new system would violate IHL. In case a legal review identifies a use that might be problematic, doctrine and training are recommended to be drafted and implemented in such a way that they allow the weapon to be used properly and in compliance with IHL.
During the R&D process, systems should be conceived in a way that provides operators with a sufficient amount of human understanding in order to achieve adequate awareness of the situation. This would also allow to obtain an account of the reasons why the machine is suggesting or going to take a specific course of action. Additionally, the design must take into consideration a function allocation procedure that allows each step in the targeting cycle to be identified and to specifically assign it to either a human or a machine.
Machines should be extensively tested during the R&D phase. Frequent feedback from actual users, in conditions close to real-life ones during system deployment is also crucial. Validation process should also be extensive in order to ensure that machines meet the required specifications and fulfill their intended mission. Additionally, systems need to be proofed against spoofing techniques and cyber or electronic attacks. Otherwise, they couldn’t be deemed feasible from an operational perspective.
Commentaries on National Implementation of the Guiding Principles on LAWS: Submitted by Spain (6 July 2020)
