Automation bias and acceptable losses: AI targeting systems and dehumanisation of Palestinians

During Operation Swords of Iron, the Israel Defence Forces (IDF) has reportedly used AI-enabled targeting systems to strike individuals suspected of links to Hamas and Palestinian Islamic Jihad. Concerns have been raised about the compliance of these operations with International Humanitarian Law (IHL) and the potential errors of AI systems, which could contribute to civilian casualties. A significant portion of those killed during this operation have been Palestinian civilians, prompting questions about the targeting decisions made by these systems and accountability for the resulting harm. This study uses critical discourse analysis to examine the language used in interviews of former and current IDF intelligence officers conducted by Yuval Abraham, while drawing on the concept of digital dehumanisation to explore the degree to which civilian lives were considered during targeting decisions, how the potential for civilian deaths was rationalised, and what made these strikes acceptable. I argue that the use of AI-enabled targeting systems in Operation Swords of Iron has contributed to the historical dehumanisation of Palestinians by providing a mechanism for Palestinians to be reduced to data and that resulting civilian casualties, caused by strikes based on flawed or biased information, violate IHL, which requires accountability for these actions.