This paper explores the ethical implications of AI-supported decision-making in military medicine through the lens of relational ethics. It employs a comparative analysis of two fictional vignettes – a civilian triage situation during a pandemic and a military triage situation during an armed conflict – to argue that the use of AI is significantly more ethically problematic in the military domain than in civilian public health emergencies. The core difference lies in the nature of the “necessity” governing each crisis. In the civilian pandemic case, “pandemic medical necessity” is an ethically ambiguous and complex concept, subject to public discourse and epistemic underdetermination. Given this ambiguity, the pragmatic benefits of an AI-support system, such as decision-making speed, may plausibly be accepted despite the system’s potential for random, opaque information alteration. In contrast, the military context is governed by “military-medical necessity”. This standard provides a much clearer and more ridig ethical and prudential guardrail for decision-making. The paper contends that current AI systems, which lack deductive reasoning and genuine understanding of human intent, are fundamentally incapable of reliably tracking or implementing this complex, normative standard. This failure renders AI incompatible with military accountability structures. These structures rely on a relational division of agential labor and deference to authority, which presuppose that agents can recognize and be bound by rules and mutual obligations. Because an AI system cannot be a full participant in these human ethical relations, its use in military medical decision-making is categorically doubtful.