The AI Resource Curse and Human Devaluation

The resource curse describes a well-documented political economy pathology: states whose revenues derive from extractable commodities rather than from taxing a productive citizenry lose the Institutional incentive to invest in human capital. The rentier state literature — from Mahdavy through Beblawi to Ross — establishes that when governments need oil more than they need people, people become politically and economically disposable. The intelligence curse extends this framework into the age of artificial general intelligence.

As AI infrastructure — compute clusters, energy grids, data pipelines — becomes the primary driver of GDP growth, the comparative return on investment in human beings declines in precisely measurable terms. The ideological surface of this shift became visible when Sam Altman reframed the cost of AI queries against the multi-decade resource cost of developing a human, a rhetorical move that treats human beings as a production function to be benchmarked against silicon alternatives. This is not an isolated provocation; it coheres with a broader tendency among some AI builders to frame their work not as tool-making but as the instantiation of a new cognitive species — a framing that implicitly repositions Homo sapiens as legacy infrastructure.

The structural consequence is that GDP, already a contested proxy for civilizational health, becomes actively misaligned with human welfare under AI-driven growth. Nations optimizing for output will rationally defund education, public health, and social reproduction. The intelligence curse is therefore not a speculative future risk but the logical terminus of applying twentieth-century economic metrics to a production regime in which human cognition is no longer the scarce input.