Modern political economy has been founded on an implicit yet powerful assumption: human labor constitutes the primary vector of value creation and, through wages, the fundamental mechanism for the distribution of wealth. Artificial intelligence undermines this arrangement not as a simple technology of substitution, but as a systemic factor that progressively separates value, labor, and decision. This is not an abrupt rupture, but a structural drift that runs through firms, markets, and institutions, placing under strain economic categories that have held for more than a century.

The disconnection between labor and value emerges as a defining feature of the computational economy. In industrial productive systems, even in their most advanced forms, value remained anchored to a combination of labor time, applied skills, and the organization of the production process. Cognitive automation alters this balance. Systems capable of classifying, predicting, optimizing, and deciding operate without a direct counterpart in terms of measurable human labor. According to estimates by the McKinsey Global Institute, up to 60 percent of activities currently performed by qualified workers contain components that are technically automatable through advanced AI. The value generated by these systems tends to concentrate where intellectual property rights, computational infrastructures, and strategic orchestration capabilities reside, not where labor is actually performed.

This dynamic redefines the very concept of productive contribution. Cognitive automation does not eliminate labor, but reduces its decision-making centrality. Human labor is progressively shifted toward functions of supervision, validation, or exception handling, while the core of the decision-making process concentrates in models and algorithms. In large data-driven organizations, this transformation is already visible: entire operational chains remain indispensable, yet economic value and internal power accumulate around a few strategic functions that control computational systems.

The consequences for wages are structural rather than contingent. The wage polarization observed in recent years cannot be explained solely by globalization or by technological progress in a generic sense. AI accentuates a fracture between those who directly participate in the design and governance of systems and those who operate in increasingly standardized executory roles. According to OECD data, in advanced countries the share of labor income in GDP has shown a steady downward trend, while the share attributable to rents from immaterial capital and technological ownership is growing. Wages progressively lose their function as the primary mechanism for value distribution, without an equally robust institutional alternative emerging.

This separation produces significant macroeconomic effects on aggregate demand. If a growing share of value concentrates in actors with a low propensity to consume, the system’s capacity to absorb its own production weakens. The International Monetary Fund has repeatedly pointed out that rising income and wealth inequalities represent a structural brake on long-term growth. In the computational economy, productive efficiency can increase more rapidly than social demand, turning optimization into a source of systemic instability.

In this context, the rationalization described by Max Weber takes on a new configuration. Weber had grasped how the rational organization of labor, founded on calculation and predictability, could turn into an “iron cage” capable of emptying human action of meaning. Computational systems push this logic further: calculation no longer merely organizes labor, but becomes itself a producer of value. When algorithmic efficiency becomes the dominant criterion, human labor risks being perceived as a residual, necessary but economically secondary.

For firms, this transformation is not an abstract problem, but a matter of strategic governance. Technological choices directly affect wage structures, the distribution of internal power, and the sustainability of the market in which the firm operates. Companies that have pushed cognitive automation without considering systemic effects have often experienced short-term margin growth accompanied by increasing demand fragility and social tensions that are difficult to manage. By contrast, organizations that integrate AI within a broader vision of value creation and distribution show greater resilience over time.

Cognitive limits play a central role in this dynamic. AI promises to overcome them, but in reality it reorganizes them. Decision-makers tend to rely on what is measurable, modelable, and comparable, neglecting qualitative dimensions of work that do not easily enter models. This is a risk well known even in advanced management: when what is not quantified ceases to matter, the real value of the organization is progressively eroded, even in the presence of formally positive indicators.

The separation between labor and value in the computational economy is not an inevitable outcome, but a trajectory that can be governed. It does, however, require an explicit collective decision. Rethinking remuneration models, mechanisms for participation in generated value, and the role of fiscal and welfare institutions is not an ideological exercise, but a systemic necessity. Without these adjustments, the economy risks producing value without being able to transform it into cohesion and stability.

Labor, in the era of artificial intelligence, does not disappear, but changes status. From a direct foundation of value, it becomes a condition of balance for a system that, if left to a purely computational logic, tends to lose its social anchoring. Understanding this transformation means recognizing that the AI economy cannot be governed exclusively with categories inherited from the past. The invisible price of intelligence is not technological, but political and institutional, and concerns what a society decides to recognize as value worthy of being sustained over time.

Global AI Observatory