The military use of AI and the law of armed conflict: what role for government and industry?

The UK has committed itself to the responsible adoption of AI technologies for military use. While the Ministry of Defence has published detailed guidance on the ethical dimension of military AI, so far it has offered remarkably little guidance on its legal aspects, including the law of armed conflict. The purpose of this article is to call attention to this imbalance and the risks it poses. A large proportion of AI capability development is now driven from the bottom up. Yet industry is not bound to ensure that its products are compliant with the law of armed conflict. Considerable time, effort and resources may therefore be invested in the design and development of AI systems that subsequently turn out to raise substantial legal concerns. This could present the Ministry of Defence with an unpalatable choice between redesigning these systems at considerable expense or deploying them at the risk of non-compliance. To avoid this outcome, we argue that the Ministry should recognize that the legal and ethical considerations of AI are distinct, determine what exactly its legal obligations demand, communicate these requirements to industry by setting detailed expectations and proactively engage with businesses early in the AI system lifecycle.