As wealth from artificial intelligence grows, a new political fight is brewing over whether to tax AI and how to share its gains. Democratic congressional candidate Alex Bores is pushing for a You have taxeswhile a tax expert warns the policy could backfire. The debate reflects growing concern that the benefits of AI are concentrated among investors and big companies, not among workers.
“Should we tax AI?
The question arises as companies invest billions in AI systems and infrastructure. Supporters say new taxes could fund worker training, protect wages and support communities facing disruption. Critics counter that a poorly designed levy could slow useful adoption, hurt small businesses and be difficult to enforce.
Growing wealth, unequal gains
“The race for AI has made a lot of people richer… but most of these gains seem to have gone to the rich while ordinary workers don’t see their incomes increase. »
This statement reflects a broader trend in recent technology cycles. Capital holders quickly benefit from rising valuations and new products. Productivity benefits for workers tend to appear later, if at all. Economists disagree on how quickly AI will change the job market, but many expect strain on certain roles, from customer support to routine office work.
Governments have already considered similar ideas. The European Parliament debated a “Tax robots.” several years ago, but was not adopted. In 2017, South Korea reduced tax incentives for automation, a measure often described as a limited form of tax on robots. These episodes show the value of capturing some gains from automation without blocking innovation.
Bores’ arguments for an AI tax
Bores argues that AI creates new flows of wealth and that public policies should direct some of these gains to local workers and services. He presents the idea as a matter of fairness and as a way to prepare for disruption.
Supporters highlight several goals for an AI tax:
- Fund worker training and apprenticeships. Support mid-career transitions to higher paying positions.
- Strengthen the safety net. Supporting communities during rapid change.
- Profit sharing. Make sure windfall profits benefit more than a small class of investors.
Policy architects are exploring ways to target the tax at AI’s sources of value: computational, data, and software models. They also want to avoid penalties for small developers or open research.
How a tax could work
Several models are under discussion in political circles. Each has trade-offs in terms of fairness, growth and ease of application.
- Calculation or energy taxes: Charges on high-end chips or data center power consumption above defined thresholds.
- Tax on exceptional profits: A surcharge on extraordinary profits linked to AI in large companies.
- Automation offset: Employer contributions when AI replaces roles, like social charges which finance training or salary insurance.
- License Fee Model: Fees on border systems that exceed security or capacity thresholds.
States could first test targeted measures, such as reporting rules for large-scale training or nominal fees for megawatt-scale AI clusters, before broader adoption.
The expert’s reservations
A tax expert interviewed during the discussion expressed skepticism. She questions whether policymakers can identify what counts as “AI” for tax purposes, given that many tools are integrated into workflows. She also warned of the risk of offshoring if costs increase domestically.
His criticisms focus on the design and practical application:
- Measuring job losses is difficult and contested.
- Taxes on computing or energy could affect research and small businesses more than giants.
- Companies could reclassify their expenses or move their operations abroad.
- Existing tools – corporate taxes, capital gains, antitrust and training credits – could be better levers.
She suggested lawmakers first test targeted credits and transparency rules. Clear reporting on AI-driven productivity and layoffs could guide further actions.
What comes next
Several paths could evolve in parallel. Congress could consider hearings on the impact of AI on work and request studies from agencies on displacement measures. States could try pilot fees or disclosure mandates tied to large-scale trainings. International negotiations, such as those carried out by the OECD, could reduce the risks of tax arbitrage.
For voters, the question is where to place the political prism: the tools that enable AI, the profits they generate, or the outcomes in labor markets. Bores’ appeal puts that choice on the table. The caution of the tax professional highlights the issue of good precision of details.
The debate is set to intensify as AI spending increases and election season heats up. Expect proposals that start small, measure results, and adjust. The central test will be whether policies can channel AI’s gains to workers without stalling progress or pushing investment overseas.
