By Emine AL; Research Advisor - Lecturer Dr. Ahmet Sahin ATA

INTRODUCTION 

The increasing integration of AI into daily life over time has led states to exert political and legal pressure on individuals.  This situation gave rise to the concept of "AI Sovereignty." Subsequently, issues such as the ethical and moral dimensions of artificial intelligence and certain unanswered questions emerged. This article explores these threats, focusing on AI's ethical dimensions, the complex issue of accountability, and its perilous integration into military technologies concerning international humanitarian law. 

Key Words: Artificial Intelligence Sovereignty, Accountability, International Humanitarian Law 

Governance & Accountability 

People sometimes prefer to choose AI due to undetermined decisions and situations, and these are occurring in some gaps of justice in legal context. This is because there are no single, unified rules in international law. For the USA, innovation precedes regulation; meaning, rather than the state, the "voluntary commitments" of corporations dictate the rules?1i Meanwhile, the USA perceives AI as a tool for military supremacy and economic leadership. Furthermore, in China, AI is viewed as a superpower to assert state control and govern society.?ii2Hence, their objective is to secure "political control" through the utilization of AI. 

The uncertainty of rules makes the concept of liability unsolved. There are ongoing debates in international law regarding whether to grant personhood to AI. 3iii While some categorize it as a mere object, others argue that it possesses a form of personhood, and finally, there are those who posit that granting AI legal personhood could facilitate legal processes. The EU's approach is also ’Civil Law Rules on Robotics’’ mentioned and here the necessity of human supervision in autonomous systems is emphasized. ?iv 

However, in human psychology, the matter becomes even more complex. Humans' sense of control and social identity are interconnected.4v Yet, when using Generative AI, control does nearly complete without the individual. The individual issues a prompt, but the AI generates the output through its internal algorithms. Finally, if a person begins to feel entirely weakness in the face of AI, accountability problems emerge. For instance, people are occasionally forced to share their personal information.5vi The major reason for it is that the AI-supported programs want to access individual's personal data to identify the things which are their aim to enhance. Herein, it is not clear whether individuals are truly free to share their personal data. In summary, the legal obligation rests with the final decision maker. Nevertheless, it needs to discuss whether this is fair or unfair considering the examples provided. 

Bias & Privacy 

Technological advancements in the field of AI have profoundly impacted the protection of personal data and the right to privacy. 6vii This dynamic is closely associated with AI's speed in acquiring, processing, and storing information. It includes both positive and negative sides. The positive aspect is its facilitation of operations in certain sectors. The negative aspect, however, is that its usage lacks sufficient transparency and comprehensibility. Because of the negative aspects, people may have some insecurities.  

The fuel of AI is "data." Data comes from individuals. Such data can be easy to access for unknown people. Online shopping, banking transactions, and other data-driven operations expose this vulnerability. Naturally, all these attempts are not malicious. 

Substantially, the initial aim of data collection is to serve the user better. However, the core problem is the lack of a definitive answer to whether the utilized data is used for benevolent or malevolent purposes. Additionally, developments in the field of AI may have national or international political outcomes regarding the right to privacy. Indeed, systems that aggregate data derived from satellite imagery, facial recognition-supported cameras, or mobile phone location tracking can monitor individuals' movements and even predict their future actions and coordinates. 7viii Also, AI carries the potential for bias in algorithmic decision-making. 8ix For instance, Google Photos scandal in 2015; the software's inability to recognize Black faces led to their classification as 'gorillas.' 9x 

While fueling societal inequality, all these factors can trigger further discrimination against societies and individuals. At this stage, the reality that classical legal instruments are inadequate in combating the negative impacts of the virtual world on the concept of personhood comes to light. 10xi The most distinct example of this is the anonymity of the perpetrator. In classical law, to initiate a lawsuit, it is necessary to know the identity and residence of the defendant; however, the creator of videos or photographs generated in the virtual world may remain unknown. This damages the meaning of the principle of territoriality.   

Thus, categorizing data is meaningless. Because they reflect the nature of the data at the moment of collection. However, they neglect its subsequent utilization and potential transformations.11xii 

Trustworthiness & Warfare 

People tend to prefer having their workload lightened and the existence of a structure that makes decisions on their behalf. Then, the question arises as to how much one can trust AI. In relation to this topic, the study conducted by Markowitz and Hancock examines AI's perception of honesty in communication across four distinct stages. 12xiii The primary objective of the study is to determine whether the "truth-bias" possessed by humans exists within AI, and if so, to what extent. It proves that AI, much like humans but in a significantly more dominant manner, operates according to the principles of TDT. AI possesses a default assumption that the texts it interacts with are factual. This situation reveals that AI is vulnerable to disinformation and manipulation and can exhibit a more pronounced "truth bias" than humans. According to the research, the psychological infrastructure lays the foundation for the most significant risk on the battlefield.  

The moment a person trusts a machine under the assumption that it won't make errors, they lose their sense of control. Specifically, with the deployment of autonomous weapons, the protection of civilians and the right to life within the context of humanitarian law are increasingly threatened. 13xiv If an operator believes that AI makes more accurate decisions in target acquisition than a human, they will not question the probability of the machine's erring. However, autonomous weapon relies solely on encoded data; since its mission is absolute eradication, eliminating the potential enemy serves as its sole motivation.  

Some regulatory policies regarding war and artificial intelligence have been enacted by leading countries. For example, it can be stated that the EU adopts a more human-centric and risk-based model in AI governance compared to the USA. However, the EU's Artificial Intelligence Act deliberately excludes concepts such as military applications and national security. The rationale behind this is likely the result of a clash between the concepts of "normative sovereignty" and "strategic imperative." 14xv  

The reason for this is that, despite the EU pursuing the goal of an "economic and political union," military decisions remain within the sovereign purview of member states. 

 CONCLUSION 

The unchecked development of artificial intelligence has multifaceted and serious consequences, primarily including the deepening of social inequalities and the threat to global security through military applications. Since AI functions as a modern Panopticon, strong legal frameworks and international cooperation are imperative. Without them, humanity risks becoming slaves to its own technological creations, prioritizing power struggles over fundamental human rights and privacy. 

Emine Al is a Political Science and Public Administration student currently enrolled in an English preparatory program. Her academic interests focus on international relations, diplomacy, and global politics.