Atlantic Council

Agentic AI systems are being procured for military command and control (C2) under public commitments to rigorous testing and human oversight. Whether such commitments can be discharged depends on their supporting assurance case, which requires three elements: claims specifying the conditions for acceptability, evidence bearing on those claims, and an argument connecting the two. Through a structured review of 240 documented Testing and Evaluation (T E) practices, spanning eight evaluation dimensions and three lifecycle stages, we identify eight assumptions that established methods make about their test article, grouped into four clusters: system specifiability, stability, composability, and supervisability. Agentic properties weaken all eight assumptions. This erosion affects the argument connecting evidence to claims, not the claims or evidence themselves. As a result, test results may satisfy process requirements, but they do not warrant the inference from tested to fielded behavior. We derive ten assurance claims for the first three assumption clusters and assess whether current and emerging methods can address each, mapping operational consequences through five C2 scenarios. Supervisability is identified but not assessed here, since evidencing it depends on system stability results and human factors T E methods beyond the present scope. The documented record does not support broad claims about system-level behavior, but narrower claims remain recoverable in principle, contingent on mature methods: bounded mission envelopes, trajectory-grounded correctness, executable runtime constraints, and characterized run-to-run variance. Part of the evidentiary burden shifts into deployment, making the determination to field a continuing act. Where evidence cannot be generated, the residual uncertainty can be governed through defined expiry conditions and assigned ownership.

This study investigates the evolving challenges intelligence and law enforcement agencies face in countering Salafi jihadist terrorist activities across digital platforms, focusing on integrating artificial intelligence (AI) and data fusion technologies in open-source intelligence (OSINT) methodologies. Through an analysis of case studies involving ISIS and other non-state actors, the study examines how terrorist organisations adapt their digital strategies to evade detection while exploiting social media platforms, instant messaging applications and websites for recruitment, propaganda, and operational planning. The study demonstrates the necessity of developing advanced technological solutions such as the integration of AI-enhanced OSINT capabilities, combined with data fusion approaches, that can improve analysts' ability to process multilingual content, identify threat actors who use online anonymity, and detect patterns of terrorist activity across multiple network layers. However, the study also reveals critical ethical and operational challenges in balancing effective counter-terrorism measures with privacy rights and civil liberties. The study concludes that while OSINT, AI and data fusion tools can substantially augment human analysis capabilities, successful counter-terrorism efforts require a carefully calibrated approach combining technological innovation and human expertise.

What motivates state and non-state actors to sponsor proxy forces? Why do proxy forces accept support, even knowing it may come at a cost to their autonomy or legitimacy? Three causal logics explain decisions to sponsor proxies. First, sponsors often have secondary strategic objectives where alignment of interest or ideology, the pursuit of influence, or greed compels them to act but is insufficient to warrant more than limited intervention. Second, actors will work through proxies when they have constraints on direct action, to include cost sensitivity, political risk, and escalation avoidance. Third, working through proxies can be the optimal means to achieve security objectives, even for high-priority concerns and when the external actor has few constraints on action, because proxies have local knowledge or will be perceived as more legitimate than an external actor. On the other hand, proxies accept both materiel and non-materiel support to tilt the balance in their direct conflicts. At the tactical and operational levels, proxies seek sponsors who can fill gaps in their organic warfighting capabilities. At the strategic level, proxies seek sponsors who can provide political and ideological overhead, international advocacy, overhead, financing, and a greater sense of purpose. History is replete with proxy conflict, and a survey of the international system suggests proxy warfare will continue to play a prominent role in international relations. National security practitioners and security studies scholars must understand the motivations behind actors' use of proxy war strategies.