Economist Argues AI Has Ended the Era of Scarce Intelligence, Redefining Education Itself

For more than half a century, the economics of education has rested on a deceptively simple idea: schooling builds human capital, human capital raises individual productivity, and productivity commands higher wages. That logic has shaped university funding formulas, government skill policies, and the personal decisions of millions of students who treated diplomas as investments with predictable returns. But artificial intelligence is now dismantling the scarcity on which the entire framework depends. In a study published on September 12, 2026 in ECNU Review of Education, researcher Yilei Shao argues that when intelligent capability becomes abundant, cheap, and instantly available on demand, human capital theory can no longer explain what education is for or how it creates value. The paper introduces a new analytical framework it calls Intelligence Surplus, which relocates the economic purpose of education from the individual to society as a whole, and it opens a debate that reaches far beyond the economics of schooling into questions of judgment, meaning, and responsible action in human-AI societies.

The evidence for this rupture is already visible in labor markets that were once considered immune to technological displacement. Major technology companies have sharply reduced their reliance on traditional academic credentials in hiring, while generative AI systems can supply expert-level knowledge and sophisticated problem-solving on request, capabilities that previously required years of formal training and professional experience. According to the article, these developments strike at a foundational assumption of modern credentialing: that educational qualifications remain reliable indicators of genuine competence. If a machine can draft legal analyses, write functional software, and synthesize medical literature, the signaling power of the degrees once attached to those tasks weakens considerably. Shao situates this shift within the intellectual tradition of five Nobel Prize-winning scholars whose theories of human capital, skill formation, signaling, and educational returns underpinned the modern economics of education. Each of those frameworks, the paper argues, rests on assumptions that AI is now disrupting simultaneously, a convergence that has no precedent in the history of educational thought.

The study identifies five interlocking assumptions that have silently governed education systems for decades. The first is linear returns to education, the belief that each additional year of schooling yields proportionally greater economic rewards. The second is stable skill depreciation, the notion that competencies erode at a predictable pace over a working lifetime. The third is individual ownership of capability, the assumption that productive skills reside within single persons. The fourth is stable signaling through credentials, the idea that diplomas reliably communicate ability to employers. The fifth is that education generates broadly positive externalities, benefiting society beyond the individual learner. Shao contends that AI unsettles all five at once. Labor markets increasingly reward people who cross critical thresholds in human-AI collaboration rather than those who simply accumulate more schooling years, turning smooth returns into step functions. Skills now depreciate in a handful of years as AI capabilities advance. Productive capability is distributed across human-machine systems rather than contained in any one mind.

The disruption extends into territory that conventional economics has barely mapped. Credentials lose their signaling reliability precisely because AI can perform many of the tasks associated with credentialed expertise, forcing employers to reassess what a degree actually certifies. Meanwhile, educational externalities are becoming structurally more complicated. The knowledge and experience that educators and professionals contribute can now be used to train AI systems whose benefits are captured by technology platforms rather than by the people and institutions that supplied the underlying data and expertise. In economic terms, the positive spillovers of education are being privatized at scale. This asymmetry, the paper suggests, demands a new analytical category, because human capital theory was never designed to describe a world in which the productive output of learning can be extracted, aggregated, and resold by third parties. It is in response to this compounded breakdown that Shao proposes Intelligence Surplus as the new central concept for the economics of education, one suited to an era in which intelligence itself has ceased to be the binding constraint on production.

Intelligence Surplus is defined as a society’s accumulated reserve of intelligent capability beyond its basic operational needs, a pool of capacity that can be mobilized for innovation, adaptation, and institutional development. The crucial conceptual move is a shift of scale. Human capital is primarily an attribute of individuals, measured in earnings premiums and years of schooling. Intelligence Surplus, by contrast, is a societal and systemic resource, closer to infrastructure or public wealth than to personal property. To give the concept analytical texture, the author draws on Pierre Bourdieu’s influential theory of cultural capital and distinguishes three forms of Intelligence Surplus. The embodied form encompasses human capacities for collaborating with AI and navigating complex human-machine environments. The objectified form is represented by technical systems such as AI models themselves. The institutionalized form consists of governance systems, rules, standards, and accountability frameworks that shape how intelligent technologies are deployed. Treating these three forms as a coherent stock of societal intelligence reframes education’s task from producing workers to stewarding a shared resource.

If intelligence is no longer scarce, the paper asks, what has become scarce instead? Shao’s answer is a capacity he calls meaning-translation: the ability to turn unique human experiences, judgments, and values into forms that can interact productively with AI systems. To describe the people who perform this function, the article introduces the concept of the First Compressor, individuals who translate irreplaceable human experiences into representations that broader intelligent systems can understand and process. These translators occupy a strategically novel position in the economy, because their work is difficult to automate. It depends on human relationships, care, interpretation, and judgment, all of which resist the pattern-extraction methods that make machine intelligence so powerful elsewhere. The First Compressor concept also carries a critical edge: it implies that the most valuable human contribution in an AI-saturated economy is not raw analytical power, which machines now supply abundantly, but the articulation of human meaning that gives that power direction. Education systems that ignore this shift risk optimizing students for capabilities the market no longer prices.

These ideas lead to a systematic rethinking of education’s purpose. Rather than primarily preparing workers with marketable skills, education is presented as the social mechanism through which societies create, sustain, and renew their Intelligence Surplus. Shao therefore proposes understanding education as a form of intelligent infrastructure, an institution whose role is to support the creation, circulation, governance, and renewal of intelligent capabilities across society, much as energy grids support economic activity. The paper sketches this transformation at three interconnected levels. At the individual level, education should cultivate dual literacy, bringing technological literacy together with humanistic literacy so that learners can operate machines without surrendering interpretation to them. At the organizational level, schools and universities should become environments that support meaningful human-AI collaboration and lifelong learning, replacing the front-loaded degree model with continuous renewal. At the societal level, governments should build educational systems deliberately designed to function as intelligent infrastructure, coordinated with governance frameworks rather than siloed as vocational pipelines.

The paper ultimately looks beyond economic productivity to a classical philosophical question: what is education actually for? Drawing on Hannah Arendt’s distinction between labor, work, and action, Shao argues that education’s deeper purpose is to cultivate people capable of exercising judgment, caring for others, creating meaning, and acting responsibly within society. In Arendt’s terms, as machines absorb more of the routine cognitive work that constitutes labor and even skilled work, the distinctly human capacity for action, the unpredictable, value-laden engagement with other people and the public world, becomes not less important but more central to the purpose of education. This reframing has concrete stakes. If policymakers continue to measure education solely by earnings returns to credentials in a labor market where credentials no longer signal competence, they risk dismantling institutions whose real function is the maintenance of societal intelligence. If they adopt the Intelligence Surplus framework instead, education funding, curriculum design, and AI governance become part of a single policy question: how to build the infrastructure that keeps human judgment, care, and meaning-making viable in a world where answers are cheap and good questions are not.

Subject of Research: Economics of education and artificial intelligence: the shift from human capital theory to Intelligence Surplus

Article Title: When intelligence is no longer scarce, what comes after human capital?

Article References: When intelligence is no longer scarce, what comes after human capital?. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: artificial intelligence, economics of education, human capital theory, Intelligence Surplus, credentials, labor markets, ECNU Review of Education, AI literacy, human-AI collaboration, educational infrastructure, meaning-translation capacity, Hannah Arendt