AI's Promise Is Fading. The Economic Risks Are Growing

TL;DR: America's tax system was built around workers earning paychecks — not machines doing their jobs. If AI changes that equation, Washington's biggest fiscal challenge may not be spending, but where future tax revenue comes from.

We were told artificial intelligence would usher in a new era of prosperity. We were promised explosive productivity, better jobs, faster economic growth and a future where AI would handle routine work while people moved into more valuable careers.

What we’ve received has been a growing wave of layoffs, and few jobs to replace them.

Instead of using AI simply to help employees work more efficiently, many companies are using it to reduce payrolls and redirect billions of dollars toward AI infrastructure, data centers and computing power. If that trend accelerates, the consequences could extend far beyond Silicon Valley. They could fundamentally reshape the nation's tax base, consumer economy and public finances.

Oracle has become the latest symbol of AI's changing relationship with the workforce.

Earlier this year, reports emerged that the company was considering cutting 20,000 to 30,000 jobs as it struggled to finance an unprecedented expansion of AI data centers and cloud infrastructure. Analysts estimated the cuts could free up as much as $10 billion annually for AI investment.

Oracle later disclosed in its annual filing that its workforce had already fallen by roughly 21,000 employees, or about 13 percent, over the previous year. The company acknowledged that "the adoption and deployment of AI technologies across our operations have resulted, and may continue to result, in reductions to our workforce."

Oracle is far from alone in this.

Amazon has eliminated thousands of corporate positions while shifting resources toward AI. Meta, Microsoft, Dell, Intel, Cisco, Salesforce, Coinbase and numerous other technology firms have announced layoffs or hiring freezes while simultaneously increasing AI investments. According to Layoffs.fyi and multiple industry reports, well over 100,000 technology jobs have disappeared during the current wave of AI-driven restructuring.

Even more troubling is that many companies are no longer presenting layoffs as temporary cost-cutting measures. They are describing them as permanent organizational changes made possible by artificial intelligence.

Harvard Business Review recently argued that many firms are reducing headcount based less on what AI can accomplish today than on what executives expect it will eventually do. In some cases, companies may be laying off experienced employees before AI has actually proven capable of replacing them, creating the risk of expensive rehiring later.

Mass layoffs alone do not necessarily create economic catastrophe. Every recession produces layoffs, and every technological revolution has displaced some workers.

What makes AI different is that economists increasingly worry it may replace labor itself—not simply change the kind of work people perform.

That concern sits at the center of a recent RAND Corporation working paper titled Federal Revenue When AI Replaces Labor. Rather than asking whether AI will boost productivity, the researchers examined what happens if AI becomes a substitute for a significant portion of the American workforce.

Their conclusion is sobering.

RAND found that approximately 84 percent of federal revenue comes from individual income taxes and payroll taxes, with roughly two-thirds of all federal revenue directly tied to labor income. In other words, America's fiscal system depends overwhelmingly on people earning paychecks. If enough workers disappear from payrolls, the government's largest source of revenue begins disappearing with them.

The researchers stress that even if AI allows businesses to produce the same amount — or even more — with fewer workers, that does not automatically replace lost tax revenue. Corporate profits are generally taxed more lightly than wages, meaning the government collects less revenue when labor income shifts to capital income. RAND estimates corporate tax rates would have to rise dramatically in some scenarios merely to offset lost labor taxes.

The report also warns of another danger: deflation.

If AI dramatically lowers the cost of producing goods and services, prices may begin falling across much of the economy. Lower prices sound attractive, but they also reduce nominal GDP and tax collections while federal debt and interest obligations remain fixed. RAND concludes that widespread AI-driven labor displacement could make servicing the national debt substantially more difficult, particularly if displaced workers fail to find comparable employment.

The paper does not predict that economic collapse is inevitable. Instead, it serves as a stress test of America's labor-dependent tax system. But its broader warning is clear: if AI replaces workers faster than new industries can absorb them, the economic shock could resemble — or exceed — a major recession.

RAND's report is only one warning, but it is no longer a lone voice. Economists at Yale's Budget Lab, researchers at the Wharton School and Boston University, and more than 200 economists and AI experts who recently signed Stanford's "We Must Act Now" statement have all cautioned that policymakers should be thinking now about how AI will reshape jobs, tax revenue and economic growth — not after the disruption has already occurred. While there remains considerable debate over how quickly AI will transform the labor market, there is growing agreement that the nation's tax system, labor market and economic policies were built for an economy where people earn wages, not one where machines perform an ever-larger share of the work.

RAND’s findings also cast new light on America's race to build AI data centers.

Across the country, State and local governments are competing to attract billions of dollars in AI infrastructure with generous tax incentives and public subsidies. Alabama has become part of that competition.

However, the employment numbers often fail to match the headlines.

As ALPolitics.com has previously reported, roughly 80 percent of data center employment occurs during construction. Once those projects are complete, even multibillion-dollar hyperscale campuses frequently operate with only a few dozen to a few hundred permanent employees. The facilities consume enormous amounts of electricity, water and land, but they are not major long-term job creators. Instead, they exist primarily to provide the computing power that allows AI systems to automate work elsewhere.

That raises an inconvenient, very uncomfortable question.

If AI itself ultimately falls short of its economic promises in much the same way data center job projections often have, taxpayers may find themselves subsidizing infrastructure that creates relatively few permanent jobs while helping eliminate employment in other sectors.

For Alabama, this debate is more than theoretical.

The State's economy depends on workers earning wages, buying homes, shopping locally and paying income and sales taxes. If AI begins replacing large numbers of white-collar workers across the country, Alabama will feel the effects even if the jobs disappear elsewhere. Reduced consumer spending, slower business growth and weaker tax collections would ripple through every level of government.

That does not mean Alabama should reject artificial intelligence or stop recruiting technology investment.

But it does suggest lawmakers should carefully examine whether today's incentive packages reflect tomorrow's economic realities.

For years, we were told AI would create a new era of shared prosperity. The evidence emerging in 2026 tells a more complicated story. Companies are investing hundreds of billions of dollars in machines while reducing the number of people they employ. RAND's research suggests that if this trend accelerates, the nation's labor-heavy tax system could come under severe strain — and by “severe strain,“ we mean “totally collapse.”

The AI revolution will almost certainly transform the economy.

The question policymakers should now be asking is whether it will transform it for the better — or simply leave fewer Americans working, fewer taxpayers supporting government, and fewer communities receiving the broad-based prosperity they were promised.

The complete RAND working paper Federal Revenue When AI Replaces Labor is available at the embedded link above or attached below: