The Responsibility Gap in Militarized Application of AI and the Construction of Global Accountability Mechanism

In the era of rapid technological development, the application of AI in the military domain is now reshaping the nature of warfare. The unexplainability of algorithmic decision-making and autonomous weapon systems, as well as the risks associated with secure data, have sparked profound ethical and legal debates. This paper focuses on the challenge of defining war responsibility brought about by the militarization of AI, specifically the "responsibility gap" caused by the disconnection between technological autonomy and human control. By analyzing typical incidents of AI weapon misfires in the Nagorno-Karabakh conflict from 2020 to 2023 and the Russia-Ukraine war, it reveals the predicament of "black-box decision-making - fragmented responsibility" under the current international law framework. Building on previous research, it proposes a responsibility determination approach based on technological traceability and introduces the idea of using blockchain and other technologies to construct a cross-border accountability system. This study finds that the separation of decision-making levels in military AI makes traditional responsibility provisions difficult to apply, and the timestamp accuracy of combat orders needs to reach a high standard to meet the requirements of the evidence chain. Finally, this paper suggests that within the framework of existing international conventions, relevant mechanisms should be improved to enhance governance over the militarization of AI, providing a fundamental reference for research and practice in this field.