Adapting AI for Warning

Artificial intelligence (AI) is expected to revolutionize intelligence production. The partnering of AI and human analysts will be critical for continued production of actionable intelligence in an age of rapid acceleration and expanding threat domains, putting AI development in the forefront of efforts to meet emerging challenges. Strategic warning, however, is different from other analytic production efforts. Warning requires inference of intent and meaning from a given set of observables in order to effectively predict near- and mid-term actions by adversaries. The forward-looking demands inherent in strategic warning argue for an approach to AI adoption that is different from that envisioned for analytic augmentation. This article outlines a linked, semi-autonomous, and scalable system of applications that could address the unique cross-domain demands of strategic warning. It is one possible way AI might be developed and incorporated to meet the expanding and evolving needs of warning

Inference · Strategic intelligence · Warning system · AI-based Problem Solving and Planning · Anomaly Detection Techniques and Applications · Bayesian Modeling and Causal Inference