In this thesis, the role of Legal Autonomous Weapons Systems (LAWS) in the German institutional-political landscape is examined. The effort guiding the thesis is to uncover and understand how LAWS have been constructed in the German institutional-political discourse from 2017-2021. The field of LAWS is comparatively under-researched due to their recent emergence and highly contested spread as a weapon of choice. By trying to answer how securitization is taking place in German political-institutional debates surrounding LAWS, the aim is to better understand the different discursive elements of securitization in the specific German context. Building on the elements of defining the securitizing actor and referent subject, evaluating which discourse elements contribute and limit a securitization of LAWS and how suggested policy measures are tied to other discourses allows for a multileveled understanding of the issue at hand. What is clear is the need to further expand research in this field with the goal of contributing to the larger body of literature within Peace and Conflict Studies, detangling present discourses and suggesting potential policy paths. Finally, the conclusion drawn in this thesis suggests that both a technological discourse and discourse of threat of other actors contribute toward securitization of LAWS while a liberal economic discourse limits the securitization of LAWS.

Since 2009, International Relations scholars have researched the role of big advocacy groups in giving access to the Campaign to Stop Killer Robots in the United Nations Convention on Certain Conventional Weapons (CCW). To further these studies, the focus of this thesis is on the progress of negotiations for the 6-year period since the issue has been adopted, asking the question – How has the Campaign to Stop Killer Robots chosen to frame lethal autonomous weapons systems, and how successful has that framing been for the period of 2013 to 2019? I argue that advocates undertook a normative securitisation process to frame the existential threat lethal autonomous weapons systems (LAWS) pose to human beings. This argument is supported by a dual method research approach of 1) semi-structured elite interviews; and 2) qualitative content analysis of reports. The findings of this research show that, while the advocacy group has not achieved success in the form of a legally binding agreement at the CCW, they have successfully developed a process of moral stigmatization of LAWS that contributes to the creation of a new humanitarian security regime.

Since 2014, the international community has been discussing how to deal with the emergence of increasingly autonomous weapons systems under the auspices of the United Nations Convention on Certain Conventional Weapons. This case study examines the behavior of those nation-states that oppose a preventive ban on lethal autonomous weapons systems in this forum. In this regard, the concept of antipreneurship is applied to the analysis of the international meetings in order to explore the resistance patterns with which antipreneurs reject the need for normative change. Analyzing the content of related documents shows that antipreneurs block formal negotiations on a ban and deem legal weapons reviews to be a sufficient regulatory instrument. Beyond that, the antipreneurs also attempt to create a beneficial image of the relationship between artificial intelligence and international humanitarian law. This behavior contributes to force norm entrepreneurs increasingly into a defensive position and sustainably undermines their entrepreneurial demand for a preventive ban. Besides, the analysis distinguishes the resistance of antipreneurs from ambivalent behavior that cannot be clearly linked neither to the behavior of entrepreneurs nor to that of antipreneurs.

This thesis analyses States’ duty to investigate grave breaches of humanitarian law and how it applies to deep reinforcement learning autonomous weapons. It identifies basic technologic intricacies related to deep reinforcement learning and discusses what issues may arise if such software is used in weapons systems. The thesis applies a legal doctrinal method to study how the technology could frustrate the grave breaches regime and hamper States’ ability to investigate suspected incidents. Furthermore, investigative standards under humanitarian law and human rights law are examined in the context of autonomous weapons systems.  The main argument is that deep reinforcement learning algorithms create a black box that is virtually impossible to investigate and consequently causes accountability issues.