인공지능(AI) 기반자율무기체계의 윤리기준에 관한 연구: 한스 요나스의 책임이론 관점에서 = A Study on Ethical Standards for AI-Based Autonomous Weapon Systems: Form the Perspective of Hans Jonas’s Theory of Responsibility

다국어 초록 (Multilingual Abstract)

A Study on Ethical Standards for AI-Based Autonomous Weapon Systems : Form the Perspective of Hans Jonas’s Theory of Responsibility Lee Ho-seok Advisor: Prof. Jang Sang-guk, Ph.D. Department of Military Science, Graduate School of Chosun University This Study aims to propose the developmental direction of ethical standards applicable to AI-Based Autonomous Weapon Systems, taking into account the unique characteristics of warfare amid the rapid advancement of Artificial Intelligence (AI) technologies and their expanding application in the defense sector. In particular, this study is grounded in the recognition that the development and operation of autonomous weapon systems constitute not merely a technical issue, but one that is intrinsically linked to human responsibility, ethical legitimacy, and normative acceptability within the international community. To achieve this objective, the study analyzes domestic prior research as well as the defense AI Ethical Standards of the United States and the United Kingdom, employing Content Analysis as a qualitative research methodology commonly used in the social sciences. As its theoretical framework, the study adopts Hans Jonas’s Theory of Responsibility, which compensates for the limitations of traditional ethical theories. Comparative analysis is conducted with a focus on the Subject of Responsibility, the Object of Responsibility, and the Limits of Responsibility. Through this process, the study identifies limitations in existing ethical standards, particularly their tendency to remain at the level of declaratory principles and their insufficient consideration of future technological expansion and the uncertainty of the battlefield. The findings indicate that, with regard to the Subject of Responsibility, attributing ethical responsibility for autonomous weapon systems to a single actor is practically inadequate. Accordingly, the study proposes the necessity of a Human-Centered, Pluralistic, and Multi-Layered Structure of Responsibility. Concerning the Object of Responsibility, the scope of responsibility should extend beyond direct victims in combat to include societies and future generations affected by indirect and long-term consequences. With respect to the Limits of Responsibility, this study argues that such limits should not be regarded as factors that weaken ethical standards, but rather as Normative Conditions necessary for responsibility to function effectively in practice. This Study emphasizes that ethical standards for AI-Based Autonomous Weapon Systems must transcend the domains of technological regulation and risk management, and instead serve as a core normative framework for ensuring the legitimacy of weapon system deployment, compliance with International Humanitarian Law, and the protection of Human Dignity. Although this study is constrained by limited access to detailed defense-related data, the proposed developmental directions may serve as a foundational reference for the future institutionalization of defense AI Ethical Standards in the Republic of Korea Armed Forces. Ultimately, this study demonstrates academic and policy significance by arguing that ethical discourse on AI-Based Autonomous Weapon Systems should expand beyond technological feasibility and battlefield efficiency toward the preservation of Human Responsibility and Human Dignity. Keywords: Defense AI, Autonomous Weapon Systems, Ethical Standards, Theory of Responsibility, AI Ethics

목차 (Table of Contents)

• 제1장 서론 1
• 제1절 연구 배경 및 목적 1
• 제2절 연구 범위 및 방법 5
• 1. 연구 범위 5
• 2. 연구 방법 6
• 제3절 선행연구 검토 8
• 제2장 이론적 배경 및 분석의 틀 16
• 제1절 인공지능(AI) 기반 자율무기체계 16
• 1. 인공지능(AI) 16
• 2. 자율무기체계 28
• 3. 자율무기체계에 대한 쟁점과 윤리기준 적용 44
• 제2절 한스 요나스의 책임이론 67
• 1. 책임이론의 개요 67
• 2. 윤리이론과 책임이론의 관계 73
• 제3절 인공지능(AI) 윤리기준 적용 사례 77
• 1. 미국 국방 인공지능(AI) 윤리기준 77
• 2. 영국 국방 인공지능(AI) 윤리기준 80
• 제4절 분석의 틀 85
• 제3장 인공지능(AI) 기반 자율무기체계 윤리기준 분석 87
• 제1절 연구 방법 87
• 1. 내용분석 방법론 87
• 2. 연구 진행 절차 90
• 제2절 윤리기준 분석 93
• 1. 사례별 분석 결과 93
• 2. 책임의 주체 112
• 3. 책임의 대상 115
• 4. 책임의 한계 118
• 제3절 소결론 121
• 제4장 인공지능(AI) 기반 자율무기체계 윤리기준 발전 방향 126
• 제1절 책임의 주체 131
• 1. 책임 주체의 다원화 131
• 2. 최종 책임의 제도적 정립: 인간 134
• 3. 책임 주체 이행을 위한 보장 메커니즘 135
• 제2절 책임의 대상 137
• 1. 생명 보호 중심의 책임 대상 설정 137
• 2. 미래 세대와 사회환경에 대한 책임 확장 139
• 제3절 책임의 한계 142
• 1. 현실 중심의 책임 한계 설정 142
• 2. 예방적 책임을 포함한 책임 한계 규범화 143
• 제5장 결론 146
• 제1절 연구의 결과 및 함의 146
• 제2절 연구의 제한사항 및 정책적 제언 148
• 참고문헌 150