Supporting case
RAND models what happens to federal revenue when AI replaces paid work
What happened
In a report published August 25, RAND researchers Carter Price and Akshaya Suresh model two situations: a transition in which AI displaces 10% of workers and an extreme thought experiment in which machines replace all human labor. The authors vary whether displaced workers find new jobs and whether AI providers charge prices near cost or collect substantial profits. In their scenarios, taxes on corporate profits do not fully replace lost taxes on labor. If AI services are sold near cost, less of the new value appears as taxable profit; if people find jobs at similar pay, the loss is smaller. These are results under stated assumptions. The study does not predict how many jobs will disappear or estimate the probability of either scenario.
The paper shows why cheap AI and a weaker tax base can coexist. Productivity gains may reach consumers as lower prices while payroll and income-tax receipts decline. Capturing part of that surplus to fund services or social insurance would require knowing where it accrues and choosing a tax that can reach it.
Dates and implementation stage
August 25, 2026: RAND report published.
