Sam Altman sees an economy on the brink. So does Mark Cuban. And Sen. Elizabeth Warren agrees the stakes could not run higher. Yet a new analysis from tax experts pushes back hard. The tax code, they argue, needs no dramatic rewrite to survive artificial intelligence.
Calls for Radical Change Meet Skepticism
Tech leaders and lawmakers have flooded recent months with proposals. Fortune reported in April how Altman’s OpenAI and investor Vinod Khosla pushed to scrap federal income taxes for Americans earning under $100,000. They want to shift the burden from labor to capital gains and corporate profits. “They will vote for a candidate who says no taxes if you make less than $100,000,” Khosla said. OpenAI’s policy paper warned the current system was “designed for an economy in which most value was created by human labor. That economy is disappearing.”
Mark Cuban floated taxes on AI tokens or compute power. Billionaire John Arnold suggested a levy on data-center processing to fund universal basic income. Time published Warren’s case in May. She called for a wealth tax so figures like “Jeff Bezos and Sam Altman shouldn’t pay lower tax rates than the workers they fire.” Warren also backed an excise tax on data-center energy use. Bigger facilities would pay more. “AI could lead to ‘a level of wealth concentration that will break society,’” she wrote.
The Economist floated a dedicated fund to cushion workers hurt by the transition. All these voices share one premise. AI will displace millions. Payroll taxes will collapse. Something must give. But Daniel Bunn and Alex Muresianu at the Tax Foundation disagree. Their piece published Friday labels the panic premature. “AI may be a transformative technology, but that is not a good justification for throwing core principles of tax policy out the window,” they write. Those principles? Simple rules. Low rates. Broad bases. No penalties on investment.
History offers little support for panic-driven overhauls. Labor-saving inventions from the internal combustion engine to the personal computer reshaped daily life. Yet labor’s share of national income held roughly steady. The U.S. labor market proved dynamic. In 2025 alone roughly 63 million people were hired and 63 million left jobs through quits, layoffs or retirement. Disruption happens. Mass unemployment does not automatically follow.
And. Leisure could rise instead. Some sectors might shift to four-day weeks without pay cuts. Workers gain time with family. The economy adapts. But. Bad fiscal policy born of speculation solves nothing. Past technological leaps actually broadened the tax base. Early America relied on tariffs and excise taxes because collection was easy at ports and distilleries. Better technology later enabled income and consumption taxes. Progress widened the net. It did not shatter the code.
Recent weeks brought fresh voices into the fray. Andrew Yang appeared on television repeating an old theme: tax the robots, stop taxing human workers. A House member penned an op-ed in The American Prospect calling for an AI tax to slow layoffs and fund public jobs programs. IBM’s CEO urged a “Goldilocks” balance — not too much regulation, not too little. Yet none of these ideas sway Bunn and Muresianu. They see the same pattern. Assume catastrophe. Then rewrite rules. The assumptions remain unproven.
OpenAI itself pointed to Goldman Sachs data showing AI already eliminates about 16,000 U.S. jobs monthly. Younger workers feel it first. The company’s paper spoke of superintelligence and systems that “cannot be easily recalled.” Khosla predicted 80 percent of current jobs automated by 2030. Warren highlighted utility bills spiking as much as 267 percent near giant data centers. These numbers grab attention. They do not yet justify upending decades of tax doctrine.
Critics inside Silicon Valley already question motives. One scholar called OpenAI’s document “comms work to provide cover for regulatory nihilism.” Companies talk responsibility while racing ahead. Political reality adds friction. Raising capital-gains rates to ordinary income levels sparked fierce opposition before. Wealth taxes face legal and practical hurdles. A compute tax sounds clean on paper. Tracking every AI calculation invites complexity and avoidance.
So what path forward makes sense? Stick to fundamentals. Keep rates low. Avoid targeted penalties on investment in new technology. Broaden the base where possible. Let markets allocate resources. If AI truly delivers abundance, tax revenue may surprise on the upside even without new levies. If displacement proves worse than expected, policymakers retain tools — expanded safety nets, retraining, targeted relief — that do not require rewriting the entire code in advance.
The debate will not fade. Altman, Cuban, Warren and others command attention for good reason. Their warnings highlight real risks. Yet the Tax Foundation’s cool assessment deserves equal weight. Speculation about singularity should not drive fiscal policy. Evidence should. Today that evidence shows a flexible labor market and stable income shares despite past waves of automation. Tomorrow may differ. But betting the tax code on that guess risks more harm than the technology itself.
Washington has time. It should use that time to watch, measure and adjust rather than preempt with grand schemes. Because once principles bend to apocalyptic forecasts, restoring them grows difficult. The internal combustion engine did not break the tax code. Neither, the authors insist, will today’s algorithms.
WebProNews is an iEntry Publication