AI Tax and Work Protection Act
- Sponsor
- Rep. Casar, Greg [D-TX-35]
- Committees
- Ways and Means Committee (primary)
- Last action
- Aug 6, 2026
Bottom line
The bill aims to mitigate potential job displacement caused by artificial intelligence by taxing AI usage and reinvesting the revenue into public sector job creation and workforce development programs.
What it actually does
This bill establishes a new excise tax on the usage of 'foundation models' of artificial intelligence, with tax rates that increase as the national unemployment rate rises. The revenue generated from this tax is dedicated to a new trust fund, which will then fund a Work Protection Administration (WPA) within the Department of Labor. The WPA is tasked with establishing a grant program for eligible entities to create jobs in various public service sectors, including child care, education, health, infrastructure, and local news, and also provides additional funding for the Workforce Innovation and Opportunity Act.
Proponents argue
Supporters argue that this bill proactively addresses the societal impact of rapidly advancing AI technology, ensuring that the economic benefits of AI are shared to support workers and communities. It creates a dedicated funding stream for essential public services and workforce retraining, preventing widespread unemployment and fostering a just transition in the face of automation. The progressive tax structure, tied to unemployment rates, ensures higher contributions when job displacement is more critical.
Opponents contend
Critics might contend that the AI token tax could stifle innovation and investment in a nascent but critical technology sector, making U.S. companies less competitive globally. They might argue that the tax mechanism is complex and could be difficult to administer, potentially leading to unintended consequences or discouraging AI adoption. Furthermore, opponents might question the efficiency and effectiveness of a large-scale government jobs program funded by this tax.
The bill is moderately complex, requiring careful reading of technical definitions and interconnected provisions. The estimated reading time is sufficient for an informed reader to grasp its core mechanisms.
Title I, Sec. 101 (adding new Subchapter D to Chapter 36 of the Internal Revenue Code)
Excise Tax on Foundation Models
This provision imposes a new excise tax on 'covered persons' who develop, sell access to, or modify large-scale artificial intelligence 'foundation models' and generate revenue or use AI to reduce their workforce. The tax amount is the greater of two calculations: either a percentage of the fair market value of AI 'tokens' processed, or a percentage of the value of AI services/transactions. Both percentages are variable, increasing when the national unemployment rate exceeds 5 percent.
Supporters argue
Supporters argue this tax ensures that the economic gains from AI, which may lead to job displacement, contribute to a fund for workforce protection and job creation. Tying the rate to unemployment makes it responsive to economic conditions, ensuring higher contributions when job displacement is more critical.
Critics contend
Critics might argue this tax could hinder AI innovation and adoption in the US, making American companies less competitive globally. They may also raise concerns about the complexity of defining and valuing 'tokens' and 'foundation models' for tax purposes, potentially leading to compliance burdens and legal challenges.
Tradeoffs
This provision balances the need to fund workforce transition and public services against concerns about stifling technological innovation and economic competitiveness in a rapidly evolving sector.
Title II, Sec. 201-202
Establishment and Use of Trust Fund
This title establishes a dedicated trust fund within the U.S. Treasury. All revenue collected from the new AI excise tax (100 percent) will be appropriated into this fund annually, starting in fiscal year 2027. The funds in this trust fund are specifically earmarked to finance the operations of the Work Protection Administration (Title III) and the Jobs Program (Title IV).
Title III, Sec. 301-302
Establishment of Work Protection Administration (WPA)
This title mandates the Secretary of Labor to establish a new Work Protection Administration (WPA) within the Department of Labor within 90 days of the bill's enactment. The WPA will be led by a Director and will be responsible for developing and implementing the comprehensive jobs grant program detailed in Title IV, serving as the central agency for the bill's job creation efforts.
Title IV, Sec. 401-402
Establishment of Grant Program for Job Creation
This title establishes a competitive grant program, administered by the WPA Director, to award funds to eligible entities (such as state/local/tribal governments, schools, higher education institutions, and non-profits). These grants are to be used to hire employees for jobs that must be performed by natural persons and provide support for various public needs, including child care, education, health, elder care, housing, infrastructure, environmental conservation, public art, and local news. Priority is given to permanent, full-time jobs and to government entities for functions typically performed by the state or local government.
Title IV, Sec. 402(d) and (e)
Job Requirements and Nondisplacement
This provision sets strict requirements for jobs created under the grant program. Employees must receive a wage not less than the greater of federal, state, local minimum wage, or the prevailing wage. They must also receive health insurance comparable to federal employees, at least 12 weeks of paid family and medical leave, and paid sick leave (1 hour for every 30 hours worked, up to 56 hours annually). Additionally, the bill includes strong non-displacement provisions, prohibiting grant funds from being used to lay off, partially displace, or replace existing employees, or to fill positions of striking workers.
Title IV, Sec. 406
Additional Workforce Innovation and Opportunity Act (WIOA) Funding
This provision authorizes the appropriation of up to 20 percent of the funds from the AI tax trust fund to support programs and training activities under the Workforce Innovation and Opportunity Act (WIOA). This additional funding is authorized for fiscal year 2027 and each fiscal year thereafter, providing a significant boost to existing federal workforce development and training initiatives.
A 'covered transaction' subject to the AI excise tax includes the use of a foundation model by the taxpayer or a sale/license to a related party if such use 'enables or results in a reduction in the workforce of the taxpayer or of such related party.'
Section 101(b)(2)(A)(ii)
Why it matters:This is standard legislative practice for defining the scope of a tax, but the specific condition for related parties is a crucial anti-avoidance measure that prevents internal corporate restructuring from circumventing the tax.
Case for: Supporters would argue this prevents companies from using internal AI development or related-party transactions to automate jobs without incurring the tax, ensuring the policy's intent to mitigate workforce reductions is not circumvented.
Case against: Critics might argue this adds significant complexity for large corporations with multiple subsidiaries, requiring detailed tracking of workforce changes linked to internal AI use, potentially creating substantial compliance burdens and administrative overhead.
Estimated impact: This provision significantly broadens the applicability of the AI tax to internal corporate automation that leads to job reductions, potentially increasing tax revenue and ensuring a more comprehensive application of the policy.
The Secretary of the Treasury, in consultation with the Secretary of Labor, may suspend the higher tax rates (those triggered by unemployment exceeding 5 percent) if the elevated unemployment is due to a 'war, pandemic, or any other massive economic shock unrelated to the use of artificial intelligence.'
Section 101(c)
Why it matters:This is a practical carve-out to ensure the tax doesn't exacerbate economic crises unrelated to AI, but it grants significant discretionary power to the Secretaries, which could be a point of contention.
Case for: Proponents would argue this provides necessary flexibility to prevent the AI tax from becoming an additional burden during unrelated national crises, ensuring the tax remains targeted at AI-driven unemployment and does not worsen other economic downturns.
Case against: Critics might raise concerns about the broad discretionary power granted to the Secretaries to determine what constitutes an 'unrelated' economic shock, potentially leading to inconsistent application, political influence, or a lack of transparency in decision-making.
Estimated impact: This provision could significantly alter tax revenue collection during periods of high unemployment caused by non-AI factors, potentially reducing the tax burden on AI companies during such times.
Defines 'foundation model' as an artificial intelligence model that is trained on broad data, generally uses self-supervision, is trained using a quantity of computing power equal to or greater than 10^25 integer or floating-point operations (or such quantity as the Secretary determines necessary for comparable capability), and is applicable across a wide range of contexts.
Section 101(b)(7)
Why it matters:This is a crucial technical definition for determining precisely which AI systems are subject to the tax. The specific computational threshold and the Secretary's discretion to adjust it are vital for the tax's applicability and future relevance.
Case for: Supporters would argue this technical definition targets the most powerful and broadly applicable AI models, which are most likely to have significant workforce impacts, while providing essential flexibility for future technological advancements through Secretarial adjustment.
Case against: Critics might argue that setting a specific computational threshold could quickly become outdated as AI technology evolves, and granting the Secretary power to adjust it introduces uncertainty and potential for political influence over the scope of the tax.
Estimated impact: This definition directly determines the scope of AI systems subject to the excise tax, thereby influencing the tax base, the amount of revenue generated, and the competitive landscape for AI development.