Autonomous weapon systems (AWS) pose significant challenges for individual criminal responsibility under the Rome Statute. While existing modes of liability—such as perpetration, ordering, aiding and abetting, and command responsibility—can theoretically apply to AWS-related crimes, their effectiveness depends on human intent and control. When autonomous systems act unpredictably, attribution of the actus reus, mens rea, and causation becomes uncertain, creating a “responsibility gap” in which no human perpetrator can be clearly identified. This paper evaluates how the International Criminal Court could adapt existing doctrines to address this gap, including broader interpretations of perpetration-by-means, omission liability, and expanded command responsibility. It also assesses proposals for new offences, prosecutorial guidelines, and non-criminal accountability mechanisms. The analysis argues that legal innovation is essential to ensure that accountability remains human-centred as warfare becomes increasingly automated.