The rise of killer robots: The impact on Autonomous Weapon Systems (AWS) on the United States’ Nuclear Deterrence

Matthew Robert, Law (2021-02-26)

Thesis (MA)--Stellenbosch University, 2021.

Thesis

ENGLISH ABSTRACT: Technology is playing a major role in changing how states conduct warfare and much research is being focused on this. This is a broad area of study and the present research will address only one aspect of it, which is the impact of Artificial Intelligence (AI)-enabled autonomous weapon systems on a state’s nuclear deterrence. More specifically, this research’s main aim is to study how the rise of autonomous weapon systems could affect the US’s ‘second-strike’ capability. This term refers to the ability of a state to strike back in response to a first attack. The US’s nuclear deterrence could be affected in two primary ways. First, it could be affected practically, in that Autonomous Weapon Systems may effectively limit the US’s ability to strike back. Secondly, they could affect the US’s perception of its second-strike capability, meaning that the US could fear that Autonomous Weapon Systems could limit its ability to strike back, but this capability does not necessarily need to exist for this to happen. As perception is a central tenet in nuclear deterrence, the US needs only to perceive their second-strike ability to be under threat to feel insecure. The secondary aim was to see if the undermining of the US’s nuclear deterrence would lead to a potential disruption in strategic stability. Strategic stability is built on the premise of states being able to successfully deter one another; undermining this would lead to instability. This thesis further chose to contextualise this study by looking at President Trump and the role of identity politics. The aim of this secondary contextualisation was to create an understanding of why the US would pursue AI-enable autonomous weapon systems and whether Trump’s populist politics could explain why. This also allowed the study to better utilise Kaldor’s ‘New War thesis’. This sequentially allowed this study to understand the US perception of potential aggressors and its grand strategy. The reason for addressing this area of study is because AI has a huge and unknown potential to affect all aspects of the military. This creates the need to understand how such a technology could affect nuclear deterrence, which is the cornerstone of US National Strategic Security Policy. Nuclear deterrence has played a central role in protecting the US from potential aggressors since the time of the Cold War until today and potentially it will do so for the foreseeable future. Furthermore, with nuclear weapons becoming more prevalent in the global arena, there is a need to understand how autonomous weapon systems could affect them. This research used semi-structured interviews and secondary data analysis in order to gather data. The data indicated that Autonomous Weapon Systems offer huge potential to undermine the US’s nuclear deterrence in the future. They currently have significant shortcomings but with technological advancements they could have an impact on the US’s nuclear deterrence. This is because they offer the potential to undermine the US’s ability to strike back. Finally, there needs to be continuous study on how Artificial Intelligence and Autonomous Weapon Systems will affect the US’s military and international conflict. However, the current major threat that the US faces comes from the cyber domain. There needs to be further study into what type of cyberattack or intrusion justifies the use of kinetic force that may then allow a state to go to war.

AFRIKAANSE OPSOMMING: Tegnologie speel 'n belangrike rol in hoe lande se metodes van oorlogvoering verander, en baie navorsing word tans daaroor gedoen. Dit is 'n wye studieveld waarvan hierdie navorsing net een aspek sal ondersoek, naamlik die invloed wat kunsmatige-intelligensie(KI)-gedrewe outonome wapenstelsels op lande se vermoë tot kernafskrikking kan hê. Die primêre doel van hierdie navorsing is dan spesifiek om te kyk hoe die opkoms van outonome wapenstelsels die VSA se vermoë tot 'n tweede slaanaanval kan beïnvloed. Dié term verwys na die vermoë van 'n land om terug te slaan in reaksie op 'n eerste aanval. Die VSA se kernafskrikking kan hoofsaaklik op twee maniere geaffekteer word. Eerstens kan dit prakties geaffekteer word deurdat outonome wapenstelsels die VSA se terugslaanvermoë doeltreffend kan inperk. Tweedens kan die VSA se persepsie van hul vermoë om tweede slaanaanvalle te loods geaffekteer word. Hulle kan naamlik vrees dat hul vermoë om terug te slaan deur outonome wapenstelsels ingeperk kan word, maar die vermoë van sodanige stelsels hoef nie noodwendig werklik te wees om die persepsie te laat ontstaan nie. Omdat persepsie 'n sentrale aanname in kernafskrikking is, moet die VSA bloot bewus wees van die moontlikheid dat hul vermoë tot 'n tweede slaanaanval bedreig kan word om onveilig te voel. Die sekondêre doel was om te kyk of die ondermyning van die VSA se kernafskrikking sou lei tot 'n potensiële ontwrigting in strategiese stabiliteit. Strategiese stabiliteit word gebou op die aanname dat lande in staat is om mekaar suksesvol af te skrik. Om hierdie vermoë te ondermyn sou kon lei tot wêreldwye onstabiliteit. Die navorsing word verder gekontekstualiseer deur te kyk na President Trump en die rol van identiteitspolitiek om sodoende te probeer begryp waarom die VSA volhardend bly streef na die gebruik van KI-gedrewe outonome wapenstelsels, en of Trump se populistiese politiek dalk 'n verklaring hiervoor kan bied. Die kontekstualisering het dit ook moontlik gemaak vir die navorsing om Kaldor se nuweoorlogshipotese beter te kan gebruik. Die rede vir die ondersoek in hierdie veld is omdat kunsmatige intelligensie 'n groot en ongekende potensiaal het om alle aspekte van oorlogvoering te affekteer. Dit is dus nodig om te begryp hoe sodanige tegnologie kernafskrikking kan beïnvloed wat die hoeksteen van die VSA se nasionale strategiese veiligheidsbeleid vorm. Kernafskrikking het sedert die Koue Oorlog 'n sentrale rol gespeel in die VSA se pantser teen potensiële aanvallers. Dit speel vandag steeds 'n rol wat in die afsienbare toekoms waarskynlik sal voortduur. Met kernwapens wat reg oor die wêreld meer prominent raak, is daar 'n toenemende behoefte om te begryp hoe outonome wapenstelsels hulle moontlik kan affekteer. Hierdie navorsing maak gebruik van halfgestruktureerde onderhoude en sekondêre data-analise om data in te samel. Die data het aangedui dat hoewel huidige stelsels betekenisvolle tekortkominge het, tegnologiese ontwikkeling kan veroorsaak dat outonome wapenstelsels groot potensiaal het om die VSA se kernafskrikking in die toekoms te beïnvloed omdat hulle die VSA se vermoë om terug te slaan kan ondermyn. Laastens is voortgesette navorsing noodsaaklik om te weet hoe kunsmatige intelligensie en outonome wapenstelsels die VSA se militêre en internasionale konflik daar sal laat uitsien. Die belangrikste bedreiging wat tans die VSA in die gesig staar, kom egter uit die kuberruimte. Verdere studie is ook nodig om vas te stel watter soort kuberaanval of -indringing die gebruik van kinetiese slaankrag regverdig en 'n land in staat sal stel om 'n oorlog aan te voor.

Please refer to this item in SUNScholar by using the following persistent URL: http://hdl.handle.net/10019.1/109860
This item appears in the following collections: