The expansion of Artificial Intelligence in military operations, particularly systems capable of autonomously identifying, selecting, or engaging targets, raises a distinct problem in Twelver Shi‘i jurisprudence. The central issue is not an abstract ruling on technology, but the jurisprudential determination of the relevant subject, attribution of conduct, and limits governing the use of force. This study asks, assuming the legitimacy of defense or combat in a particular case, under what conditions military autonomous systems may be used and, when death or property damage occurs, to whom the act and resulting guarantee should be attributed. The research adopts an analytical-jurisprudential method and relies primarily on foundational Imami sources on jihad, non-combatant protection, necessity, direct and indirect causation, destruction of property, and guarantee. Contemporary scholarship is used only to clarify the technical subject. The study finds that technical autonomy is neither an independent ground of prohibition nor a basis for granting the system normative capacity or liability. Legitimacy depends on the target, operational environment, necessity, foreseeable risk to protected persons, and the availability of safer alternatives. Liability turns on customary attribution and the relative strength of causation within the chain of design, testing, authorization, deployment, and use. The article proposes a theory of “autonomous instrumental agency with continuing human attribution,” recognizing operational autonomy without severing jurisprudential responsibility from human actors.