<![CDATA[The development of artificial intelligence (AI) in weapons systems has transformed the decision-making process in armed conflict, particularly in the stages of target identification, target verification, estimation of losses to civilians, and the use of force. These changes raise questions about the application of the principle of proportionality in International Humanitarian Law, which essentially requires an assessment of the anticipated incidental losses to civilians and civilian objects against concrete and direct military advantages. This study aims to analyze the legal construction of the principle of proportionality in the use of AI-based weapons systems, identify the problems of its application, and formulate a reconstruction of the principle of proportionality that can guarantee humanitarian protection amidst the development of modern weapons technology. This study employs normative legal research methods combining legislative, conceptual, and case approaches. The research results show that the principle of proportionality remains applicable to the use of artificial intelligence-based weapons systems through various provisions of International Humanitarian Law, including Article 36, Article 51 paragraph (5) letter b, and Article 57 of Additional Protocol I of 1977. The applicable legal construction still faces limitations in providing operational parameters for decision-making algorithms, system uncertainty, automation bias, opacity, dynamic changes in situations, and the limits of human control in the proportionality assessment process. This research offers a Dynamic Human-Controlled Proportionality Model that puts human legal judgment, explainable decision-making, dynamic reassessment, uncertainty-based civilian protection, and traceable human responsibility as the key elements of reconstructing the principle of proportionality in the use of artificial intelligence-based weapons systems.]]>
