AI Could Create a Tax Problem—Not Just a Jobs Problem

Most of the debate over artificial intelligence and employment focuses on workers: Which jobs could disappear? Which occupations will change? How many people will need to retrain?

Economists are beginning to ask a second question that could be just as consequential: What happens to government finances if AI allows businesses to produce more with significantly fewer workers earning taxable wages?

The issue is not that artificial intelligence itself automatically creates a tax crisis. AI could increase productivity, create new industries, raise wages in some occupations and generate enormous corporate profits. All of those outcomes could create taxable income of their own.

The potential problem is more structural. The U.S. tax system was largely built around an economy in which human labor generates a large share of income. Workers earn wages, employers run payroll, and governments collect income and payroll taxes. If a meaningful portion of economic output eventually shifts from labor toward machines, software and capital, economic growth could continue while some of the tax base governments currently depend on grows much more slowly.

Marketplace explored that possibility this week through an unlikely historical example: Akron, Ohio, and the collapse of the city's once-dominant rubber industry.

Akron's experience does not tell us that AI will produce the same outcome nationally. It does show what can happen when jobs supporting an entire local economy disappear faster than the institutions around them can adapt.

The U.S. Tax System Depends Heavily on Work

To understand the concern, it helps to look at where federal revenue actually comes from.

The Congressional Budget Office projects the federal government will collect approximately $5.6 trillion in revenue in fiscal 2026. Individual income taxes are expected to generate about $2.8 trillion, while payroll taxes contribute another roughly $1.8 trillion. Corporate income taxes, by comparison, are projected at about $404 billion.

Not every dollar of individual income tax comes from wages—investment income, retirement distributions, business income and capital gains matter too. But CBO explicitly notes that wages and salaries make up most of the tax base for both individual income taxes and payroll taxes.

That makes employment central to government finance in a way that is easy to overlook.

A worker earns $80,000, and part of that income becomes federal income tax. Social Security and Medicare taxes are withheld from the paycheck. The employer pays its share of payroll taxes. Depending on where the person lives and works, the same paycheck can also generate state and local income-tax revenue.

Multiply that process across more than 160 million workers and labor becomes one of the government's largest revenue engines.

AI creates a potential challenge if technology changes the relationship between economic output and taxable payroll.

A business might eventually produce the same amount—or substantially more—with fewer employees. The business may become more profitable, and its owners or shareholders may become wealthier, but that does not necessarily mean governments collect the same amount of revenue through the same channels.

Social Security Makes the Labor Connection Particularly Clear

Social Security offers perhaps the most direct example.

The program receives most of its income from payroll taxes paid by workers, employers and the self-employed. Employees and employers currently each contribute 6.2% of covered earnings, for a combined Social Security payroll tax of 12.4%, up to the 2026 taxable maximum of $184,500. Medicare also collects payroll taxes, without the same earnings cap.

The Social Security Trustees describe payroll taxes as the program's primary source of income.

That structure works best when a large share of national income arrives in the form of taxable wages.

If AI mostly helps workers become more productive and those productivity gains eventually show up in higher wages, payroll-tax revenue could remain strong or even increase.

But imagine a different outcome: companies become dramatically more productive while employing fewer people, and a larger portion of the resulting income flows to corporate profits, business owners and capital rather than wages.

Economic output could rise without payroll-tax receipts increasing proportionally.

That is the scenario tax economists are beginning to worry about.

Akron Shows What Happens When Jobs and the Tax Base Disappear Together

Marketplace uses Akron because the city has already experienced a concentrated version of this problem.

For much of the 20th century, Akron was synonymous with rubber and tire manufacturing. Goodyear, Firestone, B.F. Goodrich and General Tire helped transform it into one of America's major industrial cities. Akron's population surged from about 69,000 in 1910 to more than 208,000 by 1920 and eventually peaked near 290,000 in 1960.

Then the industry changed.

Foreign competition, technological shifts and the move toward radial tires transformed tire manufacturing. Akron's factories struggled to adapt, production moved elsewhere, and by 1982 not a single passenger tire was being manufactured in the city, according to Akron's own planning history.

The lost factories did not only eliminate manufacturing jobs.

Marketplace reports that as the city's employment base deteriorated, municipal finances fell with it. Akron ultimately had to eliminate hundreds of city jobs, reduce its vehicle fleet and borrow millions to support municipal operations.

That is the part of the story relevant to AI.

When a major employer disappears, the government does not simply lose tax payments from the company. It can lose taxes generated by thousands of paychecks. Workers spend less at local businesses. Some residents leave. Property demand can weaken. Businesses serving those workers can shrink as well.

At the same time, the community may need more government support because unemployment and economic hardship are rising.

Revenue falls just as demand for public resources increases.

AI Could Create a Similar Mismatch at a Much Larger Scale

Akron was heavily dependent on one industry in one geographic area. Artificial intelligence is different because it can potentially affect tasks across many industries at the same time.

Software development, marketing, accounting, customer service, law, finance, administrative support, research, media and other knowledge-intensive occupations are all experimenting with AI systems capable of performing portions of work that previously required human labor.

That does not mean those occupations disappear.

Technology has repeatedly changed jobs without eliminating work altogether. Computers removed some occupations while creating others. The internet destroyed some business models and produced entirely new industries. AI could similarly increase demand for workers who build, manage, supervise or complement the technology.

The fiscal concern emerges under a more specific scenario: AI increases output substantially faster than it increases labor income.

University of Virginia economist Lee Lockwood told Marketplace that a large long-term decline in labor's share of income would create problems for tax systems that rely heavily on taxing labor. Nobel laureate Joseph Stiglitz similarly argued that a tax system designed around labor would eventually require reform if AI dramatically changes where income is generated.

That is a conditional argument, not a forecast.

But it raises a useful question policymakers have rarely needed to confront at this scale: Should the tax system care whether $1 million of economic output is generated by 15 workers or one worker supported by AI?

Under today's system, the answer can be yes.

Productivity Can Rise While Wage-Tax Revenue Falls

Consider a simplified example.

Suppose a business currently employs 100 people and pays $10 million in total annual wages.

Those wages generate payroll taxes and potentially federal, state and local income taxes. Employees also spend their earnings throughout the economy.

Now imagine AI allows the company to produce more while eventually operating with 70 employees and $8 million of payroll.

The business could be healthier than ever. Its revenue could rise. Its profit margins could increase. Its value could appreciate.

But the amount of taxable payroll associated with the company has fallen.

Some of the government's lost labor-tax revenue could be replaced through higher corporate taxes, taxes on investment income or taxes generated elsewhere by the growing economy. Whether that fully occurs depends on who receives the additional income, how it is taxed and what companies do with their profits.

That distinction between productivity and labor income is at the center of the emerging AI tax debate.

An economy can become more productive without every source of tax revenue growing at the same rate.

The Problem Could Be Even More Immediate for Cities

Federal finances receive most of the attention because of Social Security and Medicare, but some local governments could feel labor-market changes faster.

State and municipal tax systems vary enormously. Some rely heavily on property taxes, others on sales taxes, and others collect significant revenue from individual income or payroll.

Akron today, for example, has a 2.5% municipal income tax, with employers withholding taxes on wages subject to the city's system.

That is not evidence that the same tax structure caused Akron's fiscal problems during the rubber collapse decades ago. It illustrates why the location of jobs still matters to many municipal budgets today.

If a major employer replaces hundreds or thousands of local workers with technology, the community could lose taxable payroll even if the company remains physically present and highly profitable.

The surrounding economic effects can deepen the impact. Fewer workers may mean fewer commuters buying lunch, fewer households buying homes, lower demand for nearby businesses and, in severe cases, population loss.

A company can become more productive while the community surrounding it becomes fiscally weaker.

That is one reason AI's economic geography may eventually matter as much as its national employment totals.

Government Could Face More Demand at the Same Time Revenue Weakens

A major technological displacement would also create pressure on the spending side of government budgets.

Workers who lose employment may qualify for unemployment insurance or other assistance. Lower-income households may become eligible for food, healthcare or housing programs. Governments could face pressure to fund retraining and workforce-development programs.

Communities experiencing substantial employment disruption may also need economic-development assistance.

That creates a difficult fiscal combination:

less taxable labor income + greater demand for government support.

The problem becomes even more complicated because the federal budget already faces significant structural pressures.

The 2026 Social Security Trustees project that the program's trust funds face long-term financing shortfalls under current law. The combined OASDI program is not projected to remain solvent indefinitely without legislative changes.

That does not mean AI is currently responsible for Social Security's financing challenges. Population aging, demographics and existing benefit and tax structures are the much more immediate factors.

But a substantial future decline in taxable payroll could become another variable policymakers have to consider.

There Is Also a Much More Optimistic AI Scenario

The tax-crisis argument should not be treated as inevitable because there is another plausible outcome.

AI could make workers more productive without eliminating large numbers of them.

A financial analyst might serve more clients. A software engineer could build applications faster. A physician could spend less time on administrative work. A small business might expand because automation allows it to operate more efficiently.

If those productivity gains increase wages, create new jobs and grow businesses, the tax base could expand rather than contract.

New industries can also appear in ways that are difficult to predict beforehand.

Few policymakers in the early internet era could have accurately forecast how many people would eventually work in cloud computing, app development, digital marketing, cybersecurity, e-commerce or social media.

AI may follow a similar pattern.

That is why the real tax question is not simply “Will AI eliminate jobs?”

It is:

“How will AI change the distribution of national income among workers, companies and owners of capital?”

The tax consequences depend heavily on the answer.

If Income Moves From Labor to Capital, Tax Policy May Eventually Have to Follow

Marketplace's reporting highlights several ideas economists are beginning to discuss if AI meaningfully erodes the labor tax base.

Some proposals involve taxing AI systems or automated production more directly. Others would shift more taxation toward corporate profits, consumption or capital. More unconventional ideas involve governments taking equity positions in technology companies so the public participates financially in productivity gains.

None of these approaches is simple.

A tax on automation could discourage businesses from adopting productivity-enhancing technology. Higher corporate taxes could affect investment or create incentives for companies to shift income elsewhere. Consumption taxes can fall disproportionately on lower-income households unless they are designed with offsets. Government ownership of private companies creates its own questions about markets and governance.

And deciding what qualifies as a "robot" or taxable AI system could become increasingly difficult as artificial intelligence becomes embedded inside ordinary business software.

The emerging debate is therefore less about identifying one obvious new tax.

It is about whether a tax system built around the economic structure of the 20th century would still work effectively in an economy where human labor and machine-generated output are divided very differently.

The National Debt Makes Waiting for a Crisis More Difficult

Marketplace's report also connects the AI question with another fiscal issue: federal debt.

Economists discussing an AI-related employment shock point out that governments generally have more flexibility to respond to crises when their finances are stronger beforehand. If a major technology transition required temporary income support, retraining, regional assistance or other intervention, borrowing could help cushion the adjustment.

But the United States is entering the AI era with federal debt already above $40 trillion and annual deficits running near $2 trillion.

That does not prevent Washington from responding to another economic disruption.

It does make every new fiscal challenge part of an already difficult budget debate.

An AI tax problem, if one eventually materializes, would therefore not arrive in isolation. It would intersect with Social Security financing, Medicare, federal interest costs, state budgets and existing disputes over taxation and spending.

This Is Why AI's Economic Impact Is Bigger Than the Jobs Number

It is tempting to measure artificial intelligence primarily by counting how many jobs it creates or eliminates.

That may eventually prove too narrow.

A worker represents more than labor to the economy. A paycheck is income for a household, a tax base for governments, funding for Social Security and Medicare, purchasing power for businesses and often the basis for retirement savings, healthcare and credit decisions.

When employment changes, all of those relationships can change with it.

Akron offers a historical example of how quickly the effects can compound when an industry supporting a large share of a community disappears.

AI is unlikely to recreate Akron's story exactly. The technology may create more jobs than expected, complement rather than replace many workers, or distribute productivity gains widely enough that government revenue remains strong.

But the underlying fiscal question deserves attention before anyone knows which outcome wins.

If technology allows the economy to generate substantially more value with substantially less taxable labor, governments may eventually have to reconsider how they finance the programs and services built around the assumption that most economic income begins with a paycheck.

That is a much bigger conversation than whether AI takes someone's job.

It is about what happens to the financial system around work when work itself changes.

Help Clients Prepare for a Financial World Where Work Is Changing

Technology can change income, employment and entire industries faster than a household's financial obligations change with them. Mortgages, bills, debt, insurance and long-term goals still have to be managed even when the way people earn money is evolving.

Copiafy gives financial professionals one AI-powered client financial workspace for organizing income, credit, debt, goals, documents and other financial information—helping you see how changes in a client's work or income could affect the rest of their financial picture.

When the economy changes, having that broader view can make it easier to identify where clients may need more flexibility and what deserves attention next.

Explore Copiafy and see how a connected client financial workspace can support your practice.

This article is provided for educational and informational purposes only and does not constitute personalized financial, tax, legal, employment or investment advice.