Even Economists Aren't Sure What AI Will Do to the Job Market | Chicago Booth Review

TEXT START: Despite widespread predictions of increased unemployment, experts are uncertain how and where jobs will be affected.

The Dissection

The article converts a structural threat into a forecasting problem. It inventories economists’ uncertainty about timing, diffusion, worker adaptation, policy, and job creation, then presents that uncertainty as the central conclusion. The result is a respectable waiting room: AI may transform labor, but nobody supposedly knows enough to diagnose the patient.

The Core Fallacy

It treats uncertainty about the path as uncertainty about the mechanism. Under the Discontinuity Thesis, the decisive question is not whether every occupation disappears or whether replacement jobs appear. It is whether AI achieves durable cost and performance superiority across cognitive work, whether institutions can preserve human-only economic domains, and whether most people retain economically necessary labor.

The article also uses the unemployment rate as its primary instrument, which is too blunt for the event being analyzed. A worker can remain technically employed while losing bargaining power, earnings, status, hours, and productive necessity. “New jobs” do not restore the mass employment-to-wage-to-consumption circuit if they are scarce residual niches, low-value service roles, or positions controlled by AI-capital owners.

Hidden Assumptions

• That historical technological transitions remain a reliable guide despite AI targeting cognitive labor across nearly every sector.
• That job creation is a sufficient counterweight to job destruction, regardless of scale, timing, wages, or ownership.
• That worker adaptation can occur at the speed required by AI deployment.
• That public policy can preserve mass productive participation rather than merely subsidize consumption after participation has collapsed.
• That uncertainty among economists is evidence against discontinuity, rather than evidence that their inherited models lack the variables required to see it.
• That “employment” still means economic necessity, when firms can retain nominal jobs while making human labor optional.

Social Function

Primarily ideological anesthetic and prestige signaling, with a partial truth embedded inside it. The uncertainty is real at the level of timing, sectoral sequencing, policy response, and distribution. But the article turns that legitimate uncertainty into permission to defer structural judgment. It gives institutions a clean alibi: wait for more data while ownership, deployment, and bargaining power move ahead of public comprehension.

The Verdict

This is not an analysis of whether the postwar labor system survives. It is an economist’s fog report from inside a model built for incremental automation. The article correctly observes that the transition’s details are uncertain, but mistakes uncertainty about the calendar for uncertainty about the direction: if AI severs productive participation for the majority, the wage-consumption circuit dies regardless of how confidently economists describe the wreckage afterward.