How Federal Tax Policy Can Address AI - ITEP.org

TEXT START: Lawmakers are increasingly being called upon to use the tools at their disposal — particularly tax policy — to address the public’s concerns about artificial intelligence (AI).

The Dissection

This brief is a fiscal containment memo. It translates a potentially civilizational rupture into three familiar policy categories: redistribute AI wealth, tax AI’s harms, and build a public stake in AI companies. Its most important admission is also its most dangerous one: AI could make most people economically obsolete.

The brief then routes that possibility through a policy toolkit designed for inequality, pollution, and corporate tax avoidance. It identifies who may capture the gains and who may bear the costs, but it does not fully confront what happens when the wage-to-consumption circuit itself is severed. Its disclaimer that it does not recommend a specific approach also maps the policy perimeter without committing to a theory of control.

The Core Fallacy

The central error is a category error: treating the collapse of productive participation as primarily a tax-incidence problem.

If cognitive automation achieves durable cost and performance superiority, tax policy can seize and redistribute part of the resulting surplus. It cannot recreate the demand for human labor that generated wages, bargaining power, social status, and mass consumption under the post-WWII order.

An automation tax treats job destruction as a negative externality. Under the Discontinuity Thesis, job destruction is not a side effect. It is the competitive function of the technology. Taxing tokens or data-center energy may price resource use, but it does not restore human necessity. Repealing full expensing may slow deployment at the margin, but it cannot permanently preserve human-only economic domains while rival firms and states continue automating.

A public stake changes ownership of the machine. It does not make the majority necessary to operate it. Redistribution can preserve consumption; it cannot preserve productive participation. The brief’s proposals are therefore lag defenses and transition controls, not a mechanism for saving post-WWII capitalism.

Hidden Assumptions

The brief smuggles in several assumptions:

• AI will resemble earlier industrial revolutions closely enough that tax reform and public investment can eventually repair the social contract.
• Congress and the IRS will define AI, tokens, displaced jobs, and taxable profits precisely enough to prevent evasion and regulatory arbitrage.
• AI firms will remain sufficiently visible, domestic, profitable, and taxable for the proposed revenue to materialize.
• Revenue spent on infrastructure, education, research, or job creation will generate durable opportunities rather than finance sectors that AI also automates.
• Taxing automation will slow replacement without simply shifting production to new firms, jurisdictions, or machine-intensive competitors.
• Income redistribution can compensate for the loss of agency, status, bargaining power, and social necessity.
• Public ownership can substitute for the mass participation that gave the old system political legitimacy.

The brief recognizes some of these problems, especially the difficulty of defining AI and attributing layoffs. It does not solve the deeper problem: under coordination impossibility, stable human-only domains cannot be preserved at scale.

The Social Function

Classification: partial truth, transition management, and ideological anesthetic.

This is not pure copium. The brief correctly identifies concentrated ownership, tax avoidance, resource consumption, infrastructure strain, and the possibility of mass obsolescence. Its anesthetic function comes from converting a structural break into a menu of administratively familiar reforms. It lets policymakers discuss who should pay without answering who commands when labor no longer has credible economic leverage.

The Verdict

The brief identifies the corpse accurately enough, then proposes tax policy as embalming fluid. It may help capture surplus, fund transfers, limit environmental damage, and establish public ownership during the transition. Those are consequential functions.

But if P1, P2, and P3 hold, the proposals do not preserve the employment-wage-consumption system. They manage its remains. The real question is not whether taxation can make AI equitable. It is whether taxation can buy enough time for control of energy, logistics, maintenance, and AI capital to be reorganized before the majority becomes economically dispensable. This brief approaches that question, but stops at fiscal redistribution.