As artificial intelligence advances, it threatens not only jobs but also the federal revenue that depends on them, according to a new research brief from the RAND Corporation. Because taxes on labor generate roughly two-thirds of U.S. federal revenue, widespread AI displacement of workers — especially highly paid professionals — could create a significant fiscal challenge.
RAND researchers Carter C. Price, Monique Martineau, and Jeffrey B. Wenger developed a framework to assess how AI-driven job displacement might affect federal tax revenue. They estimate that AI could displace 10 to 15 percent of the hours Americans work over the next 15 years, with the highest-paid workers — including lawyers, analysts, software developers, and financial professionals — facing the greatest impact.
The researchers modeled four scenarios based on two key variables: whether displaced workers find new jobs, and whether AI is priced monopolistically or at cost. The scenarios represent extreme endpoints of possible outcomes, according to RAND.
In two of the four scenarios — where displaced workers do not find new employment — federal revenue falls by 1.3 to 1.6 percentage points of gross domestic product. In the other two, where workers do find new jobs, revenue rises modestly. The study assumes a 10-percent shock to employment in each scenario.
The analysis draws on 2021 Internal Revenue Service data showing that the top 10 percent of earners earn about half of all pretax income and contribute 76 percent of federal revenue from individual income taxes — roughly one-third of all federal revenue. Displacing even 10 percent of high earners would create a fiscal shock comparable to a recession, the researchers said.
RAND also identified structural features of the current tax code that could worsen the problem. Employers pay a 7.65 percent payroll tax for each human worker, but not for software, robots, or AI systems. Depreciation rules allow businesses to write off automation investments faster than the assets wear out, effectively subsidizing worker replacement. Additionally, capital income — dividends and investment gains — is taxed at a maximum rate of 20 percent, while wages can be taxed as high as 37 percent. This gap could shift income from workers to AI asset owners and lower the overall effective tax rate, the researchers noted.
To stabilize the tax base, RAND suggests two broad reform approaches. One is replacing the employer-side payroll tax with a corporate income tax, which would eliminate the financial penalty on hiring. The other is a progressive corporate tax structured around per-worker profits, where firms with many employees relative to profits pay a lower rate than AI-intensive firms with few employees and large profits.
The research was funded by the Diller-von Furstenberg Family Foundation, Pershing Square Philanthropies, and the Nick and Leslie Hanauer Foundation. The full brief is available on the RAND website.