Artificial intelligence can’t be touched or seen, but it’s shaking up global markets and soaking up investment at a historic pace. Big Tech is pouring staggering sums into AI development, fueled by record profits. But there’s another side to the story: massive workforce reductions. Companies like Amazon, Meta, and UPS are cutting jobs as automation accelerates — and that ripple effect may soon reach public finances.
If fewer humans are working, fewer people are paying taxes. So a once-theoretical question is back in the spotlight: when machines and algorithms replace workers, should they also shoulder the taxes those workers once paid?
The Tax Puzzle Behind Automation
Labor income — via income tax and social security contributions — is a central pillar of tax systems worldwide. And the fear that automation will shrink the tax base isn’t new. Nobel laureate Edmund Phelps floated the idea of a “robot tax” back in 2019. Bill Gates suggested robots should be taxed like the workers they replace.
Sanjay Patnaik of the Brookings Institution notes that in the U.S., roughly 85% of federal tax revenue comes from labor income. If AI reduces that base, governments may face real fiscal pressure. His solution: raise capital gains taxes, rather than creating a separate AI tax, which he says would be nearly impossible to design effectively.
The truth is, no one knows yet how generative AI will play out. Forecasts vary wildly:
-
Goldman Sachs predicts AI could boost global GDP by 7% over the next decade.
-
The IMF estimates up to 0.8 percentage points of annual global growth through 2030.
-
The International Labour Organization says one in four workers could see their jobs affected — but expects most jobs to change rather than disappear.
“We know there will be an impact, but it’s difficult to quantify,” says labor law professor Luz Rodríguez. She notes that earlier waves of automation hit mid-skill jobs, while generative AI targets more skilled roles. Still, she points out that new technology also creates entirely new categories of work — think content moderators or Bitcoin miners.
Why an AI Tax Might Fall Apart in Practice
Some economists argue that a tax on AI is simply unworkable. Daniel Waldenström of the Stockholm Institute for Industrial Economics points out that despite rapid tech adoption, unemployment hasn’t risen significantly. And how would such a tax even be defined?
“What counts as AI — a chip, an app, a robot, a line of code?” he asks. “We will never be able to define it precisely.”
The IMF leans the same way: targeting AI directly could hinder innovation and distort markets. Instead, it recommends:
-
Raising capital taxes
-
Applying supplementary taxes on “excessive” corporate profits
-
Reworking R&D incentives that may unintentionally push companies toward labor-replacing technologies
Oxford’s Carl Frey agrees. The tax system, he says, has drifted out of balance: many OECD countries now tax labor more heavily while easing taxes on capital, encouraging automation over job-creating innovation.
Tech Giants Are Booming, Workers Are Not
Recent data reinforces these concerns.
-
Amazon posted a 38% jump in profits and poured billions into AI — while cutting 14,000 jobs.
-
Across the OECD, the average corporate tax rate fell from 33% in 2000 to 25% today.
-
Meanwhile, the tax burden on workers barely budged.
Robotics industry leaders push back on the idea of taxing automation. Susanne Bieller of the International Federation of Robotics argues that robots boost productivity and create jobs — and taxing tools rather than profits would hurt competitiveness. “Robots don’t take over entire jobs,” she notes. “They take over tasks.”
Beyond Jobs: The Inequality Question
Even if AI creates new jobs, the transition may not be smooth. Tech giants’ soaring valuations raise fears of another bubble, and the enormous energy demands of AI systems carry real climate risks.
In the best scenario, AI could drive the creation of better-paid, more productive, and more accessible work, offsetting losses in tax revenue. But there’s a real risk that job creation will lag behind job displacement, widening gaps between workers, sectors, and even countries.
MIT economists Daron Acemoğlu and Simon Johnson have been warning about this for years: automation has enriched companies, but shared prosperity hasn’t followed.
As Rodríguez puts it, “Technology has political consequences. We can’t simply accept technological determinism. The debate is necessary — and the future depends on the choices we make.”