Legal evaluation of ai-based object recognition systems utilizing big data analytics in the defense industry

ABSTRACT In today's defense industry, artificial intelligence and big data-driven object recognition systems play a critical role in various domains, ranging from warfare strategies to cybersecurity. However, the legal status, data protection implications, ethical dimensions, and liability frameworks governing these systems remain ambiguous. The limited human intervention in the decision-making processes of AI-based systems has raised significant debates concerning their legal and criminal liability. This thesis presents a comprehensive legal analysis of AI-driven object recognition systems in the defense industry, focusing on the Turkish Code of Obligations (TBK), Turkish Penal Code (TCK), Product Safety and Technical Regulations Law, European Union Artificial Intelligence Regulations, the Turkish Personal Data Protection Law (KVKK), and the General Data Protection Regulation (GDPR). From a contractual and tort liability perspective, the legal framework governing damages caused by artificial intelligence systems remains uncertain under TBK. This study examines the liability regime of AI-based object recognition systems in light of tort law, contractual breaches, and strict liability principles. The thesis analyzes the distribution of legal responsibilities between manufacturers, users, and administrative bodies in cases of foreseeable damages, while proposing insurance and compensation fund mechanisms to address unforeseeable damages. AI-based object recognition systems are actively utilized in automated decision-making processes, including target detection and the coordination of military operations. One of the key issues in criminal liability is who should be held accountable for offenses committed by such systems—a subject of ongoing debate in international law. Additionally, the legal interpretation of intent, negligence, and gross negligence in crimes caused by autonomous systems remains a critical area of examination. This thesis investigates the impact of human intervention in AI decision-making, analyzing whether developers, operators, or military decision-makers should bear criminal liability. Furthermore, it examines current approaches to the liability of autonomous weapon systems in international humanitarian law. A fundamental question in this context is whether AI-driven military systems qualify as "safe products" under the Product Safety and Technical Regulations Law. This research explores testing protocols, security standards, and certification mechanisms necessary for ensuring compliance and operational safety in AI-based object recognition systems. The European Union (EU) has introduced the "Artificial Intelligence Act" to establish a legal framework for AI systems. However, to what extent are these regulations applicable to military AI systems? This thesis analyzes the EU's definition of "high-risk AI," its applicability to military artificial intelligence systems, and its implications for the Turkish defense industry. It also argues that Turkey should prioritize national security by developing its own regulatory framework instead of directly adopting EU regulations. AI-driven object recognition systems pose significant risks to personal data security. How can facial recognition systems, intelligence analyses, and big data repositories used in the defense industry be aligned with KVKK and GDPR regulations? This study evaluates the impact of KVKK (Law No. 6698) on personal data processing, GDPR principles such as the "right to be forgotten" and "transparency," and their compatibility with AI systems. It also examines the legal risks and obligations arising from big data-driven AI surveillance systems. Using semi-structured interviews, literature reviews, and normative legal methodology, this thesis offers practical legal solutions for defining the legal status of artificial intelligence systems in the defense industry. The field research and expert consultations conducted as part of this study underscore the need for clarifying the liability regime, establishing explicit criminal responsibility provisions, and enhancing regulatory oversight of AI-driven military systems. The study specifically examines high-risk AI systems in defense, addressing issues related to ethical principles, data security, and biometric data processing, such as facial recognition. It identifies critical gaps in the existing legal framework and proposes the following policy recommendations: ● Define clear liability criteria for foreseeable damages and establish compensation funds for unforeseeable damages. ● Update the criminal liability framework to address artificial intelligence systems involved in war crimes or human rights violations. ● Develop a dedicated legal framework for the defense industry, rather than directly adopting EU AI regulations. ● Ensure compliance of big data-driven object recognition systems with KVKK and GDPR, introducing special data protection regulations where necessary. ● Introduce a new regulatory framework titled "AI Security and Regulation" to establish clear legal boundaries for AI applications in the defense sector. This study provides a comprehensive legal perspective on the use of AI in the defense industry, identifying criminal and civil liability implications while proposing concrete regulatory measures for the integration of AI-driven object recognition systems into the legal framework. Keywords: Defense industry, big data, artificial intelligence, object recognition systems, legal liability, criminal liability, product safety, EU AI regulations, KVKK, GDPR, data security, regulatory technologies.