by Vivaldo Moscatelli
Generative artificial intelligence has now ceased to be an experimental frontier, rapidly and pervasively entering our processes of design, communication, and decision-making. Faced with this transformation, the reaction of the educational world today oscillates between two specular and equally dangerous tendencies: on the one hand, a purely instrumental enthusiasm that reduces AI to a trivial repertoire of replicable “tricks”; on the other, a defensive resistance that dismisses technology as a threat to human creativity.
Both of these positions fail because they ignore a fundamental truth: foundational models are not simple tools at our disposal, but true Complex Adaptive Systems (CAS). They are non-deterministic entities that continuously learn and react back upon the communicative ecosystems that feed them, creating a human-machine continuum in constant retraining. We cannot hope to govern this complexity with an instruction manual. From this epistemological and methodological urgency emerges the CompleXAI LAB, a project designed to guide professionals beyond the “instant solution” logic typical of traditional training courses. The beating heart of this laboratory lies in the brilliant integration between complexity science and the five stages of skill acquisition theorized by Dreyfus. The path outlines a clear evolutionary map: it begins with the Novice stage, where the user applies abstract rules within an asymmetric relationship with the machine, and progresses toward the Expert stage, in which the operator guides the interaction through situated, pre-reflective, and intuitive judgments. This model marks the end of mechanical training and inaugurates a true process of “cognitive co-evolution.” While the artificial system recalibrates its own patterns, the operator learns to govern the machine through prompting strategies and critical verification. Each module of the CompleXAI LAB is structured around a rigorous cycle: operational practice, critical reflection, and generalization. This alternation becomes the lever that accelerates deep learning, transforming declarative knowledge into tacit competence. The objective is not to accumulate rules, but to refine human sensitivity to the machine’s signals, learning to “play” with technology and accepting the margin of indeterminacy as an extraordinary cognitive resource. In summary, the goal is to cultivate a relational intelligence capable of dwelling in uncertainty, achieving a dynamic and vital balance between algorithmic structure and human intuition. The success of this methodology, initially tested on groups of trainers and consultants, now opens a new and decisive frontier of research-action: the CompleXHR LAB. This evolution shifts the focus from general education to the vital ecosystem of Human Resources, an area in which the interdependencies between algorithmic decisions, data governance, and workplace cultures enormously amplify the ethical and managerial stakes. The aim of the new laboratory is to outline advanced HR practices capable of firmly preserving human agency while still embracing the extraordinary exploratory capacity of the machine. We are at a turning point: training for GenAI means overcoming the old dichotomy between a tool to exploit and a threat to defend against, and moving toward the use of AI as a sparring partner, because the era of cognitive automation demands sovereign minds. Artificial Intelligence must be conceived as a formidable cognitive partner within a conscious process of professional maturation — an ally that does not replace our thinking, but instead pushes us, through a healthy friction, to exercise the one competence that makes us irreplaceable: high human discretion.
