Can AI behave ethically during military crises? Preserving human moral agency

Abstract This article examines the ethical implications of integrating artificial intelligence-powered decision-support systems (AI-DSS) into strategic decision-making. As AI-DSS enhances battlefield awareness and accelerates operational tempo, it also risks undermining human moral agency through automation bias, anthropomorphism and over-reliance on machine outputs. Focusing on the socio-technical and psychological dimensions of human-AI interaction, the study explores how these systems may reshape ethical deliberation, responsibility and judgement in high-stakes environments. To address these challenges, the article proposes a hybrid ethical framework that integrates elements of virtue ethics with deontological constraints and consequentialist reasoning. This approach aims to support—not supplant—human decision-making by embedding ethical considerations into the design and deployment of AI-DSS. Key contributions of the article include the adaptation of ethical exemplarism to military applications, the articulation of normative design criteria for AI-DSS and the development of policy-relevant safeguards such as mandatory override mechanisms, trust calibration systems and context-specific training. The findings suggest that AI-DSS should be understood not as passive tools but as active participants in an evolving decision ecology, raising urgent questions for strategy, policy and the governance of autonomous warfare.