Culture Is Infrastructure

The institutions that governed creative economies for three centuries—copyright law, collecting societies, arts councils, public broadcasters—were designed for a world of discrete works, identifiable copies, and traceable distribution. AI training dissolves all three. The creative output of the open web has been absorbed into model weights through a mechanism that is not copying but enclosure: value is captured as statistical patterns distributed across billions of parameters. The institutional architecture cannot process this mechanism because the assumptions it was designed for no longer exist.

Institutions are coordination architectures. Their capacity for change is measured by adaptive bandwidth (B)—how fast they can revise their settings when technology, markets, and governance fall out of alignment. When technology advances faster than institutions can adapt, an architecture gap (Δ) opens. In cultural infrastructure, B is measurably below the rate of change, and the gap is widening. The sector is in the premature regime.

Industrial policy for the intelligence age that skips cultural infrastructure is building on sand. Economic systems rest on the shared symbolic architecture through which communities make meaning, propagate memory, and survive crises. The question is not whether AI will transform culture, as it already has. The question is whether institutional architectures with sufficient bandwidth will exist to coordinate that transformation.