Adeshina SOWEMIMO
| Abstract: | Autonomous targeting systems are becoming more common and established as an armament system, yet international humanitarian law (IHL) was designed for human decision-makers, not algorithms. Focusing on the 2023-24 Gaza conflict, this study questions the effectiveness of existing legal frameworks applied during a conflict in which reports indicate AI systems, such as Lavender, Habsora, and Where’s Daddy?, were used to generate targets at unprecedented levels and speeds in modern warfare. Based on Critical Algorithm Studies and socio-technical assemblage theory, and using a legal-analytical, case-based approach, the study investigates the meaning and implications of IHL principles of distinction, proportionality, and precaution in an algorithmic context, and their failure. It found that the existing frameworks are structurally deficient because the operational conditions of AI-assisted targeting systematically undermine the cognitive and institutional practices that underpin compliance. This phenomenon, digital dehumanization, in which people are treated as part of a probabilistic data classification rather than as legal persons – widens compliance gaps and, in certain cases, leads to new legal breaches. The study concludes with concrete governance proposals, such as auditability standards, disaggregated liability-attribution frameworks, and minimum human-verification protocols, as the basis for an AI-specific legal regime suitable to contemporary algorithmic warfare. |
| Keywords: | AI-assisted targeting, Digital dehumanization, International Humanitarian Law (IHL), Lethal Autonomous Weapon Systems (LAWS) |