This paper explores the evolving dual-use potential of artificial intelligence (AI) at the intersection of military and medical domains, emphasizing the unique ethical challenges this convergence presents. Drawing on a “trimodal property model,” the author examines how AI’s application in both sectors creates shared vulnerabilities and implementation hurdles, including data misuse, automation bias, and blurred boundaries between civilian and military uses. Central to the discussion are the normative tensions between human dignity, security imperatives, and transparency, especially in military medical contexts. Hähnel proposes two practical tools—Accurate Risk Classification (ARC) and Problem Mirroring (PM)—to assess and mitigate both direct and higher-level dual-use risks. The paper argues for a hybrid, human-centered ethics-by-design approach and calls for ethics education as a critical component of AI deployment in military medicine.