The discourse surrounding Artificial Intelligence and the future of work often oscillates between two extremes: techno-optimism and doomsday predictions of a "jobs apocalypse." A recent report from MIT highlights a growing consensus within Silicon Valley that significant labor displacement is inevitable, with leaders like Anthropic’s Dario Amodei suggesting AI could become a "general labor substitute" within five years. However, University of Chicago economist Alex Imas argues that our current tools for predicting this shift are fundamentally flawed.

This review examines the shift from measuring "AI exposure" to understanding "price elasticity" as the critical metric for determining the survival of human professions.

The Problem with "Exposure"

To date, the primary method for assessing AI’s impact on the workforce has been the analysis of task exposure. By breaking jobs down into individual tasks, a methodology utilized by both OpenAI and Anthropic, researchers estimate what percentage of a role can be automated.

Imas posits that exposure is a "meaningless tool for predicting displacement." While a high exposure rating (such as the 28% cited for real estate agents) indicates that AI can assist in a job, it does not dictate whether the human worker becomes obsolete or more valuable. The "elevator operator" scenario, where a job disappears because AI can perform 100% of the tasks more cheaply, is a rare edge case. For the rest of the economy, the outcome depends on a more complex economic lever: demand.

The Elasticity Argument

The core of Imas’s argument rests on price elasticity of demand. When AI makes a worker more productive, the cost of the service or product they provide typically drops. The future of that job then depends on how the market responds to that lower price:

1. Elastic Demand: If a price drop leads to a massive surge in demand (e.g., more people buying premium apps because they are now affordable), the company may actually hire more workers to handle the scale of growth, despite each worker being more efficient.
2. Inelastic Demand: If demand remains stagnant regardless of price (e.g., a service people only need once, regardless of cost), the increased efficiency leads directly to layoffs, as fewer humans are needed to meet the fixed demand.

The Data Gap: A "Manhattan Project" for Economics

The most striking takeaway from the MIT report is the revelation that we lack the data necessary to make these predictions. While economists have granular data on the price elasticity of consumer goods like milk or cereal, we have almost no centralized data for professional services like tutoring, legal research, or dietetics.

Imas calls for a "Manhattan Project" to collect this data across the entire economy. Without it, lawmakers are "operating in the dark," unable to articulate a coherent plan for the "breakdown of the early-career ladder" or the potential recession predicted by societal impact researchers.

Conclusion

The MIT article serves as a sobering reminder that the "AI jobs" debate is currently high on speculation but low on actionable data. By shifting the focus from what AI can do (exposure) to how the market will react (elasticity), Imas provides a roadmap for more rigorous policy planning. However, the clock is ticking; as AI models transition from simple task-completion tools to agentic systems, the window to collect this data and prepare the workforce is rapidly closing.

For policymakers, the message is clear: stop counting tasks and start measuring the market's appetite for AI-driven abundance. Only then can we move from a state of "panic" to a state of preparation.

Reference

O'Donnell, James. "The One Piece of Data That Could Actually Shed Light on Your Job and AI." MIT Technology Review, 6 Apr. 2026

The one piece of data that could actually shed light on your job and AI

“We need a Manhattan Project for this,” one economist says.

MIT Technology ReviewJames O’Donnell