This paper develops a specialised ethical framework for evaluating the use of Artificial Intelligence (AI) in intelligence operations (“intelligence-AI”). While AI enhances analytical and predictive capacities, its integration into intelligence work raises distinctive ethical concerns regarding privacy, autonomy and social cohesion. Concerns that are not addressed by mainstream AI ethics. The framework developed in this paper provides the conceptual tools to understand not only the type of harm intelligence-AI can cause but also its particular level. The framework then links these levels of harm to a corresponding justificatory scale grounded in anticipatory self-defence. This scale is calibrated by threat magnitude, temporal proximity, evidential quality and target liability, and provides a structured basis for determining if and when intelligence-AI may be permissible. This novel approach provides a more robust conceptualisation of how to resolve the ethical tension caused by using potentially harmful intelligence tools while still recognising their importance for protecting people from threats, thereby equipping intelligence operators with the tools for more insightful reflection on the material the AI produces.
Ross W. Bellaby (Mon,) studied this question.