131 6 CONCLUSION In this research we based our implementation concept on existing operational norms within the Dutch Ministry of Defense, for example Rules of Engagement, the limits of an Area of Operation and Restricted Operating Zone, which are available before the deployment of a military mission. Although we conducted a value elicitation in our empirical investigation phase, we did not translate these values into the observable norms of our implementation concept. However, we are convinced that is theoretically and practically possible to translate these identified values in observable norms, but the extent of this needs to be researched. Our sixth and final research question: Q6 To what extent can observable criteria for Human Oversight be incorporated in an implementation concept for the deployment of Autonomous Weapon Systems? was answered by designing and building an implementation concept to operationalise the Glass Box framework as an example to show that the framework is actionable. We have created he CPNs as a means to gain an understanding of the observation stage of the Glass Box framework by visualizing the processes and to check if it is possible to apply the monitoring and verification process to a practical case of autonomous surveillance drones by modelling the Pre-Flight Mission Planning Process and Post-Flight Mission Evaluation Process. The implementation concept shows that it is possible to set observable criteria in the pre-flight process and to evaluate these criteria post-flight. During flight, the drone itself is treated as a black box of which the internal logic is not accessible. Although being a toy example, it demonstrates that a monitoring process can be designed to guarantee human oversight where the users set norms - or criteria in this example - for input and observe and evaluate the output against the input to check for noncompliance of the norms. Deviations of the norms will be reported in the verification process and can be used to update the norms in a new scenario. This way the users do not need technical skills to understand the internal workings of a drone, but still can monitor and oversee the use of the autonomous system based on observable norms. We do not monitor the in-flight actions of the autonomous drone in our implementation concept, because we assume that in-flight communication is not possible, for example due to a failing communication structure, an Electronic Warfare Threat, or operator unpreparedness. Therefore, it is not possible to oversee norm violations nor is it possible to intervene during the flight as would often be the case with black box autonomous systems. Based on this part of our research we can conclude that it is possible to simulate observable criteria in an implementation concept for the deployment of Autonomous Weapon Systems.
RkJQdWJsaXNoZXIy MjY0ODMw