To understand why generative AI is creating a labor market bottleneck, we must look to the pioneering work of MIT economists Daron Acemoglu and Pascual Restrepo. Their framework models the economy not as a collection of static jobs, but as a dynamic continuum of tasks ((i \in [N-1, N])) performed by either human labor or capital.

1. Displacement Effect vs. Reinstatement Effect

Economic history demonstrates that automation exerts two countervailing forces on aggregate labor demand:

  • Displacement Effect ((\Delta D)): Capital directly replaces human labor on existing tasks, reducing labor’s share of national income.
  • Reinstatement Effect ((\Delta R)): Technological progress invents entirely new complex tasks in which humans hold a comparative advantage, expanding the task space and reviving labor demand.

Net Labor Demand Growth∝ΔR−ΔD+Productivity Effect\text{Net Labor Demand Growth} \propto \Delta R - \Delta D + \text{Productivity Effect}

The Generative Asymmetry

"In previous industrial waves, new tasks (e.g. software debugging, system administration) were accessible to junior workers with basic training. In the generative AI era, the newly created tasks—such as multi-model architecture design and prompt verification—require deep domain expertise that junior candidates do not yet possess."

📚

Verified Primary Sources & Citations

Every empirical claim, economic metric, and technical assertion in this publication is cross-referenced against primary research literature and regulatory records: