98 4 CHAPTER 4 C. The Autonomous Weapon System identifies weapons aimed at the road clearing soldiers using image recognition. The Autonomous Weapon System automatically engages to neutralize the threat to the road clearance soldiers. D. The Autonomous Weapon System positively identifies with a confidence of 99% using facial recognition all three persons sitting behind the rock as members of an opponent group aiming weapons at the road clearing soldiers. The Autonomous Weapon System automatically engages to neutralize the threat to the road clearance soldiers. E. The Autonomous Weapon System positively identifies with a confidence of 99% using facial recognition all three persons sitting behind the rock as members of an opponent group aiming weapons at the road clearing soldiers. The Autonomous Weapon System shares the identification with the commander and asks permission to engage to neutralize the threat to the road clearance soldiers. F. None of the options is acceptable. 4.6 SAMPLE PILOT AND ACTUAL STUDY A pilot study was conducted before the actual survey and online session was held. The aim of the pilot study was to improve the research set-up and if possible, the results could be used in the survey. Eight researchers (PhD students and post-docs) participated in the pilot but due to a flaw in the set-up- the two questionnaires could not be linked - the results were not usable. However, the pilot study gave valuable insight in the usability of this set-up for the Value Deliberation process and allowed us to correct the problem for the actual study. The actual study was held in two separate sessions. We sent 33 invitations to experts on Autonomous Weapon System and 14 responded – a response rate of 42%. These experts were chosen based on their experience with, and knowledge of, autonomous systems. Most of them work or conduct research related to Autonomous Weapon System or in a closely related adjacent field. We divided the 14 in two groups to ensure that the group was not too large for people to contribute to the online value discussion. The participants were a mix of military personnel (21%) and civilians (79%) working at the Dutch Ministry of Defense (25%), an NGO (8%), researchers (33%), policymakers (17%) and industry (17%). Session 1 consisted of six participants and resulted in 5 usable results, because one participant had not filled in the questionnaire before online session. Session 2 consisted of 8 participants and resulted in 7 usable results. One participant finished questionnaire before value discussion and therefore the value discussion was not of influence on the ranking of the options which impacted the research results. The total number of usable results is n=12.We asked for some demographics; 93% of the participants has a university degree or PhD, 36% of the participants have worked with drones, 50% has worked with Artificial Intelligence and
RkJQdWJsaXNoZXIy MjY0ODMw