We are observing a worrying rise in civilian harm in armed conflict and at the same time an increasing autonomy in warfare. Both phenomena are happening against the backdrop of 75 years of the Geneva Conventions and the upcoming 50th anniversary of their first two Additional Protocols. In this context, international humanitarian law (IHL), with its hallmark of balancing the principles of humanity and military necessity, is facing unprecedented pressure. Over the past three decades, the narrative has shifted toward so-called “clean” warfare through more precise munitions and methods. However, this period has also seen an unprecedented number of civilian harm incidents, largely due to an increasingly broad interpretation of legal targeting permissions—a phenomenon we call the warification of IHL. The convergence of warification and increasing battlefield autonomy has come into sharp focus in the ongoing conflict between Israel and Hamas. Reports indicate that the IDF has used artificial intelligence (AI)-enabled decision-support systems (AI-DSS), including “Lavender”, “The Gospel” and “Where’s Daddy”, to help generate and process targeting lists. These systems have played a role in the high civilian casualty numbers seen in this conflict, reaching multiple tens of thousands. In this chapter, we argue that certain uses of AI-DSS may risk exacerbating this warification and propose a path forward with the legitimacy of military operations at its core. We emphasise transparency, accountability, compliance with the rule of law (both its letter and spirit) and civilian harm mitigation as key elements of legitimacy. These elements guide the shift away from warification and toward restoring IHL’s delicate balance.