Economists Still Can't Agree What AI Will Do to Jobs

The most concrete thing that can be said about artificial intelligence and the labor market right now comes from the University of Chicago Booth School of Business: even the economists studying the question don't know the answer. That admission, from one of the most closely watched business-school research communities in the United States, is itself a data point worth taking seriously.

For workers, that uncertainty is not abstract. It affects hiring plans, wage negotiations, and career bets made this year, not in some distant future. When the professional forecasters decline to commit to a prediction, the practical implication for an individual is straightforward: treat confident claims about AI eliminating or creating jobs — from any vendor, consultancy, or LinkedIn post — as unverified.

Why the uncertainty itself matters

Labor economists have a strong track record of reaching consensus on mature questions. Headline employment reports, unemployment rates, and wage growth figures typically arrive with methodology, sample sizes, and revision schedules attached, which lets analysts separate signal from noise. AI's effect on work does not yet fit that pattern.

The Booth School's position — signaled directly in the title of its own coverage, "Even Economists Aren't Sure What AI Will Do to the Job Market" — places the question in a different category. This is not a debate where competing forecasts cluster around a narrow range. It is a debate where the underlying mechanism is still contested.

That distinction matters for anyone reading labor-market commentary. A disagreement about whether payroll growth will be 150,000 or 180,000 next month is a normal forecasting spread. A disagreement about whether a general-purpose technology will expand or contract employment overall is a disagreement about the model itself.

What this means for your market position

Translate the macro uncertainty into individual terms. If economists cannot yet establish the direction of AI's effect, then no worker can reliably price their own exposure to it. Three practical implications follow.

First, skills-based decisions carry unusually wide error bars. A training investment that looks decisive today rests on assumptions that leading researchers have not validated. Diversifying skills, rather than concentrating on a single AI-adjacent bet, hedges against both outcomes.

Second, employer claims about AI-driven restructuring deserve scrutiny. Companies are making workforce decisions under the same uncertainty the economists face. A role eliminated "because of AI" may reflect a genuine technological shift or a managerial guess — and current research cannot yet distinguish the two at scale.

Third, wage signals remain the most honest indicator available. Where compensation for a particular skill moves, that reflects real demand, not forecast. Individual workers reading actual posted salaries and hiring volumes in their field are, in a real sense, doing the same thing researchers do: waiting for the data.

The difference between direction and magnitude

Part of what keeps economists divided is that AI is a general-purpose technology. Its effects will not land uniformly across sectors, occupations, or geographies. A national-level "AI and jobs" headline conceals the variation that actually determines individual outcomes: task composition within a role, the degree to which an employer has deployed the tools, and the local supply of workers who can use them.

That variation is precisely why the macro question remains open even as anecdotal evidence accumulates in both directions. Stories of automation displacing workers and stories of AI-augmented productivity raising pay can both be true. Reconciling them requires the kind of longitudinal, role-level data that researchers are only beginning to assemble.

What to watch

The honest read of the current moment is that the burden of proof is on the data, not on the pundits. The University of Chicago Booth School of Business — an institution whose researchers have shaped how the profession measures labor-market dynamics — is telling its audience that the evidence has not yet arrived.

Until it does, workers and employers alike should expect continued revision of early claims, and should weight observable market signals — postings, salaries, hiring and layoff counts — over projections. As the research community gathers the longitudinal evidence needed to settle the question, the most defensible career strategy is the one that performs reasonably under both a displacement scenario and an augmentation scenario.