139 7 CONCLUSION there are any control gaps that need to be filled for humans to remain in control and ensure accountability over these systems. Another methodological direction for future work is validating the Comphrensive Human Oversight Framework, the implementation concept and Five-Point Systems Framework with a rigorous scientific method to review and improve them if necessary. Substantively, a research area for future work are the instruments in the blocks of the Comphrensive Human Oversight Framework. Especially the adding a governance instrument during the deployment phase is in block 2 is crucial. In the military domain the targeting process is an instrument that is used during the deployment of a weapon to govern its usage, but an oversight mechanism in block 2 seems to be missing which indicates a gap in the governance layer. As an oversight process is lacking, there is no sufficient mechanism for an institution to govern or supervise the ongoing control (block 5) of a (weapon) system in the socio-technical layer. The lack of an oversight mechanism in block 2 may lead to deficiencies in the ongoing control mechanism in block 5. A second substantive area for further research is the implementation concept as operationalisation of the Glass Box framework, which aims at monitoring the behaviour of autonomous systems during operations, for example decisions of a drone during its flight. In our implementation concept we do not monitor in-flight communication as we assume this is not possible in our scenario. In a future iteration of the implementation concept a scenario with in-flight communication should be studied to see how the safety of the system and its decisions should be monitored and documented in order to account for its behaviour if norm violations occur. Another direction for future work is extending the implementation concept to other values such as privacy (for example during information gathering nearby a village) and other human rights including more fuzzier norms. This will increase the usability of the implementation concept and enhance the accountability of the autonomous systems it monitors. 7.4 CONTRIBUTIONS The scientific contributions of this research are twofold in that (1) our research contributes to a delineation of accountability, responsibility and Human Oversight that adds to the current body of literature, (2) insight into people’s perception on accountability and responsibility during the deployment of an Autonomous Weapon Systems based on an empirical value elicitation process, and (3) the framework and implementation concept for Human Oversight for Autonomous Weapon Systems might also be applied to other fields to enhance transparency of decision-making by algorithms for Autonomous Systems, such as those for Autonomous Vehicles or in the medical domain.
RkJQdWJsaXNoZXIy MjY0ODMw