AI could funnel income to a few asset-owning firms and hollow out the labor tax base; the Article urges raising capital‑gains taxes on AI‑intensive firms now and preparing to switch to a consumption tax if labor’s share collapses.

Taxing AI

Assaf Harpaz · Fetched March 15, 2026 · Social Science Research Network

semantic_scholar commentary low evidence 7/10 relevance DOI Source

The Article contends that AI will concentrate economic returns in a small set of capital-owning firms, eroding the labor tax base and recommending higher capital gains taxes on AI-intensive firm ownership now and a contingency shift to a broad-based consumption tax if labor’s share falls materially.

Artificial intelligence (AI) is poised to transform the distribution and sources of income, with some experts predicting widespread job displacement. Even under optimistic projections, AI is expected to exacerbate wealth inequality, given that the technology’s ownership and immense value are concentrated within a subset of Big Tech companies and AI startups. These outcomes will have far-reaching impacts on the federal tax system, which heavily relies on taxing individual labor income and payroll, rather than capital or consumption. This Article argues that AI threatens to disrupt the tax system’s ability to fulfill its fundamental goals of raising revenue, redistributing income, and regulating taxpayer behavior. The risk is heightened by the federal government’s dependence on individual labor income, even as economic value shifts toward mobile capital and AI ownership by large firms. The Article proposes two interventions to help address these challenges. In the short term, it recommends increasing the capital gains rates on the sale of ownership interests in AI-intensive firms. This approach would primarily serve to internalize the distributive imbalances generated by wealth concentration in AI firms. In the long term, the Article suggests adopting a broad-based consumption tax if the share of labor income declines. This proposal would represent a more drastic change, aimed at rebalancing a tax system that can no longer depend on taxing individual labor income.

Summary

Main Finding

AI-driven shifts in production and ownership will concentrate economic value in capital and a small set of AI-owning firms, undermining a federal tax system that depends on taxing individual labor income. To preserve revenue-raising, redistribution, and regulatory functions, the Article recommends (1) in the short term, higher capital gains rates on sales of ownership in AI-intensive firms, and (2) in the long term, adoption of a broad-based consumption tax if labor’s share of income materially declines.

Key Points

AI is likely to accelerate concentration of wealth and income in a subset of Big Tech firms and AI-focused startups because ownership captures large portions of AI’s economic returns.
Even optimistic labor-market scenarios foresee an exacerbation of wealth inequality due to concentrated ownership of AI value.
The current U.S. federal tax system relies heavily on taxing individual labor income and payroll taxes; a shift toward capital- and firm-based returns threatens revenue stability and the system’s redistributive/regulatory roles.
Short-term policy response: raise capital gains tax rates on sales of ownership interests in AI-intensive firms to help internalize distributive imbalances resulting from concentrated AI ownership.
Long-term policy response: transition to a broad-based consumption tax if the labor income share falls substantially, since consumption taxation targets a more stable base when labor income shrinks.
The proposals are framed as complementary: incremental redistribution via capital gains taxation now, structural tax-base reform later if warranted by realized changes in income composition.

Data & Methods

Methodological approach: normative legal and economic policy analysis drawing on projections about AI’s economic effects and on tax-design principles (revenue, equity, behavioral incentives).
Evidence base: the Article synthesizes expert projections and theoretical arguments about labor-displacing technologies and ownership concentration. (No specific empirical dataset or econometric estimation is described in the provided text.)
Analytical tools: comparative policy reasoning about tax incidence, distributional consequences, and feasibility of tax changes (short-run adjustments vs. long-run tax-base reform).
Assumptions: significant, persistent shift of economic value from labor to capital/firm ownership; political and administrative capacity to implement targeted capital gains adjustments and, if necessary, a consumption tax.

Implications for AI Economics

Tax base and incidence: AI-driven declines in labor income will change who bears tax burdens; capital and mobile returns will become more central, complicating revenue collection and requiring reconsideration of tax instruments.
Distributional outcomes: concentrated AI ownership can increase wealth inequality; targeted capital gains taxation can mitigate some inequality by taxing realizations of firm ownership value.
Incentives and investment: higher capital gains rates on AI-intensive firms could affect risk-taking, startup formation, and investment incentives—trade-offs that require careful calibration to avoid stifling innovation.
Policy design challenges: defining “AI-intensive” firms, administratively implementing targeted capital gains rules, and transitioning to a consumption tax each pose measurement, enforcement, and political-economy hurdles.
International considerations: mobile capital and cross-border ownership mean unilateral tax increases may face avoidance or flight; coordination and anti-avoidance measures will be important.
Research needs: empirical measurement of AI intensity across firms and industries; evidence on changes in labor’s share attributable to AI; revenue and behavioral effects of higher capital gains rates targeted at AI firms; distributional modeling of a shift to consumption taxation.

Assessment

Paper Typecommentary Evidence Strengthlow — The argument is normative and synthetic rather than empirical: it relies on projections, theoretical reasoning, and analogies from prior technological change rather than causal estimation or original data documenting AI-driven shifts in income composition or concentration. Methods Rigorlow — The piece uses careful policy reasoning and tax-design principles but does not present formal models, empirical tests, or robustness checks; key claims rest on strong assumptions about future AI impacts, firm behavior, and political feasibility. SampleNo original empirical sample or dataset; the Article synthesizes expert projections, theoretical literature on technology-driven concentration and tax design principles, and illustrative policy comparisons. Themesgovernance inequality labor_markets GeneralizabilityFocus on U.S. federal tax system limits applicability to countries with different tax structures and social insurance arrangements, Relies on assumptions about the magnitude and timing of AI-driven shifts in labor vs. capital income that may not occur or may vary across sectors, Heterogeneity across industries and firm types (e.g., non-AI-intensive sectors) could alter distributional and revenue effects, Political, administrative, and international tax-response constraints (capital mobility, avoidance) may reduce implementability, Definitions and measurement of 'AI-intensive' firms are unresolved, limiting practical targeting and enforcement

Claims (9)

Claim	Direction	Confidence	Outcome	Details
Artificial intelligence (AI) is poised to transform the distribution and sources of income. Fiscal And Macroeconomic	mixed	medium	distribution and sources of income	0.02
Some experts predict widespread job displacement due to AI. Job Displacement	negative	medium	job displacement / employment levels	0.02
Even under optimistic projections, AI is expected to exacerbate wealth inequality because ownership and immense value are concentrated within a subset of Big Tech companies and AI startups. Inequality	negative	medium	wealth inequality (distribution of wealth)	0.02
These AI-driven outcomes will have far-reaching impacts on the federal tax system, which heavily relies on taxing individual labor income and payroll rather than capital or consumption. Fiscal And Macroeconomic	negative	medium	federal tax revenue composition (share from individual labor income and payroll vs. capital/consumption)	0.02
AI threatens to disrupt the tax system’s ability to fulfill its fundamental goals of raising revenue, redistributing income, and regulating taxpayer behavior. Fiscal And Macroeconomic	negative	medium	tax system performance on revenue raising, income redistribution, and behavioral regulation	0.02
The risk to the tax system is heightened by the federal government’s dependence on individual labor income even as economic value shifts toward mobile capital and AI ownership by large firms. Fiscal And Macroeconomic	negative	medium	vulnerability of tax base (share of revenue from labor income) given shifts toward capital/firm ownership	0.02
In the short term, increasing capital gains rates on the sale of ownership interests in AI-intensive firms would help internalize the distributive imbalances generated by wealth concentration in AI firms. Fiscal And Macroeconomic	positive	low	distributional impacts (wealth concentration), tax incidence from capital gains rate changes	0.01
In the long term, adopting a broad-based consumption tax should be considered if the share of labor income declines. Fiscal And Macroeconomic	positive	low	tax system balance/revenue stability as labor income share declines	0.01
A broad-based consumption tax would rebalance a tax system that can no longer depend on taxing individual labor income. Fiscal And Macroeconomic	positive	low	tax system rebalancing (reliance on consumption versus labor income for revenue)	0.01