Joint Air and Space Power has always been shaped by the relationship between people and technology. Every major advance requires new forms of human competence. Early innovations such as the autopilot began to relocate that competence from direct control to supervisory oversight. Today’s AI‑enabled systems extend that relocation into far more complex socio‑technical settings.
In earlier eras, humans could master a discrete machine whose workings were visible and whose boundaries were clear. That separation has become far less clear. As AI reshapes the socio-technical systems in which humans operate, it also shifts the moment at which human judgement can still make a decisive impact. Humans no longer work through a single machine but within a landscape of AI-generated outputs whose rate of data processing is often beyond their reach. In this environment, the decisive variable is not the performance of advanced systems alone. It is the ability of humans to understand, interpret and team with these AI‑enabled capabilities under pressure. This is not the loss of control; it is the evolution of control. It makes human relevance more dynamic, more time-sensitive and more dependent on understanding how the socio-technical system behaves in practice. With it, the human role becomes strategic.
Understanding this evolution is now a core operational competence. A persistent misconception, or fear, is that AI will take over human decision-making. The operational reality is more subtle. AI-enabled systems do not eliminate the human role; they relocate it. As automation accelerates sensing, processing and decision-support functions, the various phases in which humans can meaningfully intervene do not disappear, they shift within the system. Some move earlier, into system configuration and the curation of data. Others move later, into interpretation, override and recovery. Some become supervisory rather than direct or distributed across teams rather than concentrated on a single operator. And some become so time-compressed that the human role shifts from acting to shaping the conditions under which action occurs.
To operate effectively in this environment, air and space forces need a broader concept of competence, one that enables practitioners to understand not only what the machine is doing but how the entire socio-technical system behaves around it. This is the essence of system literacy: understanding which aspects of the socio-technical system matter for effective, responsible and ethical performance.
The shift is easiest to see in practice. In AI-enabled targeting workflows, systems now generate detections, classifications and confidence scores faster than any human can scan raw feeds. The operator’s role moves from finding to interpreting and validating machine cues and shaping the upstream conditions, including data quality, sensor configuration, rulesets, that determine what the system fundamentally perceives. In air and missile defence AI generates predicted trajectories and engagement windows by integrating radars, passive sensors, command nodes and engagement platforms; the human role moves from calculating within one weapon system to evaluating predictions emerging from a layered, networked architecture. In multi-domain command and control operators no longer manage each feed directly but manage tempo, prioritisation and meaning across machine-generated summaries, controlling not the system but how the system is used.
In much of today’s defence discourse, system literacy is treated as an attribute of technical expertise: the operator’s ability to understand how an algorithm works, how data flows through a system, or how to interpret machine-generated recommendations. Useful, but incomplete.
Technology does not act alone. It acts within procedures, training pipelines, staffing models, workarounds, expectations and institutional habits. These dynamics both enable what technology can do and limit how it can be used. From the socio-technical perspective, system literacy is therefore something far broader. It involves not only understanding the technical aspects but also how technology behaves with and within the social system around it. It means seeing technology as part of a wider configuration, encompassing the organisational structures in which it is embedded, the incentives and constraints shaping its use, the informal practices that emerge around it, the cultural assumptions guiding interpretation and the institutional pressures bearing on decisions, and as one element in the system-of-systems context in which modern capabilities increasingly operate.
System literacy involves the competence to recognise how technologies, organisations and practices shape one another. It requires the operator to understand human-machine teaming as a socio-technical configuration created through the interplay of people, technologies, procedures, norms and organisational histories. To be literate in this context is to see the system in its entirety rather than the machine alone.
Although system literacy is expressed through teams, workflows, and command philosophies, it ultimately depends on individuals who understand how actions, tools and organisational context shape one another. This competence is collective, but it originates in people who can see how human and machine contributions interlock in practice.
Developing this competence allows individuals to recognise shifting windows of opportunity, to understand how machine behaviour interacts with organisational routine, to anticipate where intervention remains possible and to act in ways that elevate the performance of the whole system.
Those systems have their own tendencies. They compress, generalise and infer; they accelerate and abstract; and they will often produce an answer even when their internal confidence is low. These are not flaws but features of how contemporary AI works, and they become operationally meaningful only in combination with the social dynamics that shape how outputs are interpreted, trusted or acted upon.
Traditional training pipelines focus on platform mastery, procedural competence and tactical decision-making. These remain essential. System literacy, however, requires additional content: how AI systems learn and fail; how data quality shapes operational outcomes; how automation interacts with human cognition; how to interrogate machine outputs under uncertainty; and how to detect under-trust and over-trust in machine behaviour before it shapes decisions. These topics are not about technology alone. They capture how people and technologies shape one another in practice and are therefore operational necessities.
The solution for NATO air and space forces is not simply to add new modules to existing curricula. It is to recognise that AI-enabled systems change the nature of work itself and that training must evolve accordingly.
System literacy does not arise from classroom instruction alone. It develops through exposure to the real conditions of work: ambiguity, time constraints, organisational constraints and the gap between design and behaviour. It also depends on institutional choices: what organisations choose to emphasise, what they reward and what they allow to vary locally.
This is the cognitive edge that this year’s theme identifies as a frontier in its own right. Air and space forces that invest in system literacy, not as a technical add-on but as a socio-technical competence, will be better positioned to harness AI-enabled capabilities without losing the judgement, creativity and ethical grounding that define military professionalism.
The question is not whether AI will transform Joint Air and Space Power. It will. The question is whether NATO can develop the system literacy required to ensure that transformation strengthens, rather than weakens, the Alliance’s ability to act with speed, coherence and judgement.
