House Democrats propose AI tax to fund worker protections
Washington is starting to legislate the AI labor question directly: House Democrats unveiled the first bill to tax AI companies on usage, while a fresh PwC survey shows most CEOs still can't point to returns from the technology.
House Democrats introduced a bill that would tax AI companies on tokens and AI product revenue to fund a federal jobs program. The AI Tax and Work Protection Act, unveiled Thursday by Reps. Greg Casar (D-TX), Valerie Foushee (D-NC) and Sara Jacobs (D-CA), would impose a levy calculated on the higher of two amounts: the value of the tokens a company sells or the revenue it generates from AI products. The rate would rise and fall with the national unemployment rate, and for open-weight models the tax falls on whichever company deploys the model to cut workforce costs. Revenue would bankroll a new Work Protection Administration, modeled on FDR's Works Progress Administration, that would create jobs in housing, infrastructure, child care and elder care through grants to states, cities, tribes, nonprofits, unions and schools.
"We are not going to let AI company CEOs get rich by displacing millions of American workers," Casar said in announcing the bill, which also aims to cancel out the tax break companies currently receive when they automate a job away. The bill arrives amid escalating warnings — Anthropic's Dario Amodei has predicted unemployment not seen since the Great Depression — and it is far from the only Democratic proposal in the pipeline: Sen. Ron Wyden is pushing a data-center tax and Sen. Elizabeth Warren has floated an energy-consumption levy, per NOTUS. The Casar bill has essentially no path through a Republican-controlled House, but it matters as a marker: it is the first serious federal attempt to tax AI usage itself, and it forces the question of whether the people displaced by automation — not just the labs building it — get a share of the winnings.
PwC's survey of 4,454 CEOs found that 56 percent still see no measurable revenue or cost benefit from AI, while only 12 percent report both. The consulting giant's 29th Global CEO Survey, which polled leaders across 95 countries, found the companies that did see returns were roughly three times more likely to have their foundations in order first — governance, systems integration, and data the models could actually use. The technology itself was rarely the variable; what it was plugged into was.
That finding sharpens the backdrop for the tax bill above: most of corporate America is still paying for AI without seeing it pay back, which is precisely the kind of gap that fuels both worker anxiety and legislative intervention. For companies, the lesson is unglamorous — the models were never the bottleneck; the plumbing around them was.
What to watch: whether Casar's bill gets a hearing before the midterms, and how the token-based tax definition holds up under scrutiny from tax lawyers who have already started poking holes in it.
Do you think taxing AI usage is the right way to fund worker protections — or does it just tax the technology before it delivers? Tell us in the comments.
