The transformation produced by artificial intelligence affects the economic system at a point that precedes the labor market and measured productivity: the relationship between learning, professional continuity, and economic value. Every advanced economy has historically been founded on an implicit promise of dynamic stability, according to which investment in education allows individuals to maintain over time a recognizable position within the productive system. The large-scale adoption of intelligent systems undermines this promise, introducing a structural discontinuity between learning and durability.
The first element of rupture is the accelerated obsolescence of skills. According to estimates by the World Economic Forum, by 2027 more than 40 percent of the skills currently considered central in mid-level professional roles will be radically transformed or incorporated into automated systems. This is not a simple rotation of knowledge, but its internalization within computational models. Analytical, procedural, and decision-making capabilities are absorbed by systems that learn continuously, reducing the need for equivalent human learning in terms of depth and stability. Knowledge is not surpassed, but rendered invisible.
In this context, continuous training emerges as the dominant response, but quickly reveals its structural limits. The idea of constantly updating skills presupposes the existence of a sufficiently long horizon of use to justify the investment in training. Artificial intelligence, by contrast, drastically shortens this horizon. OECD data indicate that in several high-technology sectors the relevance cycle of a technical skill has been reduced to less than three years. Training thus tends to shift from a strategic investment to a tactical activity, necessary but incapable of producing cumulative advantages.
This shift has direct effects on the structure of human capital. Traditionally understood as a set of skills that can be accumulated and valorized over time, human capital enters into crisis when accumulation loses economic meaning. Firms experience a growing paradox: investments in training increase, but the ability to stabilize roles, careers, and expectations declines. Knowledge does not sediment, but circulates in an unstable form, often becoming obsolete before it is fully operational.
The fragmentation of professional trajectories is a systemic consequence of this process. Artificial intelligence does not eliminate qualified work, but makes its relevance intermittent. Careers unfold as sequences of adaptation rather than as recognizable progressions. What is lost is not only employment security, but the possibility of constructing a coherent professional identity over time. In macroeconomic terms, this instability reduces the effectiveness of educational investments and weakens the link between the education system and the productive system.
In this perspective, the thought of Ivan Illich takes on renewed relevance. His critique of the institutionalization of learning anticipates a risk today amplified by AI: the transformation of education into a ritual of permanent adaptation to systems that evolve more rapidly than the human capacity to internalize them. Learning risks losing its emancipatory function and becoming a reactive response, devoid of horizon. It does not produce autonomy, but dependence on ever shorter cycles of updating.
The corporate world represents a privileged observatory of this fracture. Decisions about which skills to develop and for how long become choices with high strategic risk. According to a McKinsey survey from 2024, more than 60 percent of executives believe that a significant portion of upskilling programs launched over the past five years has not generated lasting value. Not due to pedagogical inefficacy, but because technological change occurred faster than expected. Leadership is thus called not to build careers, but to orchestrate continuous transitions, with rising costs in terms of coordination and trust.
Cognitive limits come into play at a higher level. Automation promises to compensate for them, but makes it more complex to anticipate which skills will retain value. The difficulty is not technical, but systemic. In the absence of stable trajectories, learning loses its function as an economic and symbolic anchor. This produces significant social effects: disengagement from education, reduction of long-term cognitive investment, and a growing gap between individual expectations and real possibilities.
The emerging crisis of human capital is not a crisis of talent, but of structure. The problem does not lie in a lack of skills, but in their inability to translate into economic and institutional continuity. Artificial intelligence does not destroy education, but puts its cumulative paradigm into crisis. It shifts attention from competence as possession to competence as a temporary relationship with evolving systems.
In this framework, learning no longer coincides with accumulating knowledge, but with developing the capacity to orient oneself within unstable contexts. It is a meta-cognitive competence, difficult to certify and poorly recognized by traditional mechanisms of economic valorization. Human capital does not disappear, but enters a phase of suspension, in which the value of learning depends less on the content acquired and more on the ability to inhabit uncertainty without dissolving into it.
The transformation underway opens a field of inquiry concerning the relationship between education, work, and collective responsibility. In a computational economy, training can no longer mean merely updating skills, but redefining the conditions of legitimacy of cognitive investment. In this unstable space, the question of human capital remains open, not due to a lack of technical tools, but because of the absence of a new economic and institutional pact capable of restoring to learning a horizon that is not purely transitory.
Global AI Observatory