Endogenous Technological Change
Summary
Romer models long-run growth as the result of purposeful investment in new ideas. Because designs are nonrival but partly excludable, they can be reused across production while still generating market power; in the model, human capital allocated to research helps determine the growth rate.
Why it matters
If AI changes the productivity, scale, or cost of research, it may alter the rate at which economically useful ideas are produced rather than deliver only a one-time productivity gain. The framework also connects knowledge creation to ownership, research incentives, and growth.
What to keep in mind
The paper is a foundational model, not an estimate of AI’s effect on growth. Strong acceleration claims require additional assumptions about returns to research, bottlenecks, market structure, and whether machine-generated ideas can be verified and implemented.