The increasing use of artificial intelligence (AI) in military decision-support systems has raised important legal questions concerning the foundations of state responsibility. Rather than serving merely as technical tools, these systems now assist military commanders in analysing data, identifying targets, prioritising operational options, and generating recommendations that may influence decisions to use force. As a result, military decisions may no longer reflect purely independent human judgement but instead emerge from the interaction between commanders, algorithmic systems, software developers, data providers, and private contractors. This article examines whether the existing rules of state responsibility, particularly those governing the attribution of internationally wrongful acts, remain adequate to address harm arising from AI-assisted military decision-making. It argues that, although the final decision formally remains with the commander, reliance on algorithmic recommendations may significantly reduce meaningful human oversight, particularly in operational environments characterised by time pressure and large volumes of information. This concern is reinforced by automation bias, whereby decision-makers tend to place undue confidence in automated recommendations. The article concludes that the involvement of AI does not relieve states of international responsibility, since algorithmic systems lack international legal personality. However, the traditional rules of attribution require a more nuanced interpretation to address challenges relating to explainability, transparency, automation bias, and the growing role of private contractors. It therefore proposes a composite legal standard based on meaningful human control, traceability,