The Knowledge Industry

In the AI Story

The knowledge industry framing rested on a specific institutional compact — articulated by Vannevar Bush in 1945 and institutionalized by Kerr's multiversity. The government would fund basic research without directing it; the university would produce knowledge that eventually, through pathways no one could predict, would generate economic and military advantage. The compact was simple, the returns extraordinary, and the bargain held for half a century because no alternative existed. No corporation could justify funding decades of basic research whose commercial applications were unknowable; no government agency could produce the talent pipeline the research required; only the university could do both at once.

AI transforms the production process so fundamentally that the compact's terms require renegotiation. The graduate student of 2025 can produce more research, faster, across a wider range of questions than her 2005 predecessor could have imagined — large language models compress literature review from months to hours, augment experimental design, accelerate analysis, speed manuscript preparation. The answering machinery has acquired an extraordinarily powerful new production facility. But production capacity is not production quality, and quality in research depends on the quality of the questions asked — a capacity AI does not originate.

The industry connection is also transforming. Corporate laboratories — Google DeepMind, OpenAI, Anthropic, Meta AI — now produce AI research at a scale and speed that university departments cannot match. The talent pipeline has reversed: instead of the university training researchers who move to industry, industry recruits researchers away from universities before their academic training completes. The frontier AI labs have budgets and computing resources that make competing for the top talent structurally impossible for most universities. The knowledge industry still exists, but its center of gravity has shifted.

What remains for the university is the function no corporate laboratory can replicate — not the production of deployable products, but the cultivation of judgment that determines which questions deserve asking. The knowledge industry's next era depends not on whether universities can out-produce corporations (they cannot) but on whether they can out-curate them — whether the institution that was designed to produce knowledge can retool to produce the evaluative capacity that determines whether knowledge is worth producing.

Origin

Kerr borrowed the phrase from economist Fritz Machlup, whose 1962 The Production and Distribution of Knowledge in the United States had quantified the knowledge sector's economic significance. Kerr elevated it from descriptive economics to institutional self-understanding in the 1963 Godkin Lectures, where it became the frame through which the research university understood its own social function for the remainder of the twentieth century.

Key Ideas

Knowledge as economic engine. Kerr's claim that knowledge production had become the decisive factor in national economic growth, replacing earlier drivers like transportation and heavy industry.

Federal compact. The postwar arrangement — government funds, universities produce, industry consumes — that sustained fifty years of basic research.

Production versus curation. AI accelerates production but cannot originate the questions that determine what is worth producing.

Corporate competition. Frontier AI labs now produce foundational research at scales universities cannot match, restructuring the industry's talent geography.

Curator's function. The university's surviving role is evaluative — determining which research deserves pursuit when answers are cheap.