In this report, the authors analyze how artificial intelligence (AI) replacing human labor could affect U.S. federal revenue. If displaced workers do not find new jobs, and especially if AI is widely accessible at cost, federal revenue could fall sharply. Corporate profits from AI would not fully offset lost labor taxes. The authors stress the need for proactive fiscal policy to address these risks and maintain economic stability.

Federal Revenue When AI Replaces Labor

An Examination of Economic Scenarios with Highly Capable Artificial Intelligence

Carter C. Price, Akshaya Suresh

ResearchPublished Aug 25, 2026

In this report, the authors examine the fiscal implications for the U.S. federal government in scenarios in which artificial intelligence (AI) and automation technologies replace human labor. Using a framework that considers whether displaced workers find new jobs and how AI is priced (monopolistically or at cost), they analyze both a hypothetical case of full displacement of human labor and a transitional scenario in which AI displaces 10 percent of the workforce.

Simulation results indicate that federal revenue is highly vulnerable to labor-displacing AI. The loss of high-income jobs would disproportionately affect federal revenue, and increased corporate profits from AI would not fully offset declines in labor-derived taxes unless corporate tax rates are significantly raised. If AI is widely accessible at cost, deflation and reduced corporate profits could further erode tax receipts.

The authors highlight the need for proactive fiscal policy adjustments to mitigate risks from transformative AI, emphasizing the importance of monitoring economic indicators and preparing now for potential disruptions to federal revenue and economic stability.

Key Takeaways

Labor-displacing technologies could disrupt the majority of federal revenue

  • In 2024, 84 percent of federal revenue came from either individual taxes or payroll taxes, both of which are collected from labor income.
  • Because of the concentration of income at the top of the distribution in high-paid, white-collar jobs, the replacement of these roles by AI would have an outsize effect on federal revenue.
  • The loss of jobs and income would reduce economic activity in general because most households rely on labor income to support consumption. This reduction in economic activity could adversely affect businesses, which would further reduce federal and state revenues.
  • Even if displaced workers seek reemployment and find positions with similar compensation, the fiscal shock would be equivalent to a recession.

Increased revenue from corporate taxes stemming from increased profits from AI would not offset the decline in revenue from personal income taxes

  • If AI systems are owned primarily by monopolists, the majority of wages and salaries from the labor replaced by AI systems would become profit for businesses. However, corporate tax rates would need to be roughly doubled to be at parity with labor tax rates.
  • Low-cost AI could substantially reduce federal revenue stemming from business and capital. If highly capable AI systems are widely accessible at or near the cost of inference, many firms could be replaced or undercut by AI-provided goods and services with little or no profit margin. The resulting AI-induced deflation could reduce tax receipts from corporate income.

Recommendations

  • Change the tax code to reduce the dependence on labor income. Shifting revenue collection toward corporate and excise taxes and eliminating tax avoidance schemes, such as moving intellectual property to foreign subsidiaries, could reduce the fiscal shock from labor-displacing AI.
  • Stabilize nominal demand if AI causes deflation. If low-cost AI leads to broad price declines, nominal gross domestic product and federal revenue could fall even if real output remains stable. Monetary or fiscal tools could help prevent a deflationary spiral but should be used cautiously to avoid increasing the debt burden.
  • Plan income support if labor income falls sharply. A universal income or transfer program could help sustain consumption if AI substantially reduces demand for labor, as in the two scenarios in which displaced workers do not find new jobs. Such a program would require a robust funding source to avoid worsening fiscal pressures.
  • Regulate concentrated AI systems if windfall profits emerge. If labor-replacing AI produces large profits for a small number of firms, policymakers could consider regulation, special taxation, or nationalization to manage the returns from windfall profits and reduce the disruption to the labor market.

Topics

Document Details

Citation

Chicago Manual of Style

Price, Carter C. and Akshaya Suresh, Federal Revenue When AI Replaces Labor: An Examination of Economic Scenarios with Highly Capable Artificial Intelligence. Santa Monica, CA: RAND Corporation, 2026. https://www.rand.org/pubs/research_reports/RRA4980-1.html.
BibTeX RIS

Research conducted by

This work was supported by the Diller-von Furstenberg Family Foundation and Pershing Square Philanthropies and conducted by the RAND von Furstenberg Family Budget Model Initiative within RAND Education, Employment, and Infrastructure.

This publication is part of the RAND research report series. Research reports present research findings and objective analysis that address the challenges facing the public and private sectors. All RAND research reports undergo rigorous peer review to ensure high standards for research quality and objectivity.

This document and trademark(s) contained herein are protected by law. This representation of RAND intellectual property is provided for noncommercial use only. Unauthorized posting of this publication online is prohibited; linking directly to this product page is encouraged. Permission is required from RAND to reproduce, or reuse in another form, any of its research documents for commercial purposes. For information on reprint and reuse permissions, please visit www.rand.org/pubs/permissions.

RAND is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.

Version Note

This publication supersedes a previous version published in 2025 (WR-A4443-1).