AI Job Losses Haven't Shown Up in the Data Yet, But Democrats and Economists Already Want a Tax on It

Bernie Sanders wants to know how the government lets AI "throw millions of people out on the street" without a plan. He said it at a press conference in April. Since then, House Democrats Greg Casar, Valerie Foushee, and Sara Jacobs have floated a jobs program funded by a tax on AI companies, according to Fox News.

The mass unemployment hasn't happened.

The Bureau of Labor Statistics data cited by Fox News shows the U.S. has kept adding jobs every year since AI chatbots went mainstream in late 2022. Unemployment has stayed below 4.5% the entire time. The Wall Street Journal, also cited by Fox News, reported that some companies actually resumed hiring junior workers after realizing AI wasn't replacing them the way they expected.

So Democrats are proposing a solution to a crisis the labor market isn't currently showing.

The Economist Making the Same Case, From a Different Angle

Former Reserve Bank of India governor Raghuram Rajan isn't a Democratic activist, but he's landed on a similar policy idea from an economics angle. In an August 14 column for Project Syndicate, later republished by BasisPointInsight.com, Rajan proposed taxing the AI "tokens" companies use to replace human labor.

His logic: when a company hires a person, it pays payroll taxes and social security contributions. When it deploys AI instead, there's no equivalent charge. Rajan argues that gap can make automation look cheaper than it actually is, once you account for the wider cost of job displacement.

"A US firm contributes social-security payments for every worker, but not for AI," Rajan wrote.

His proposed fix: start with a low tax on AI usage, raise it gradually as data comes in, and include foreign AI providers so companies can't dodge it by routing through offshore models.

Rajan's Own Numbers Undercut the Urgency

Rajan's own data, cited identically by NDTV Profit, inkl, and whalesbook, shows AI adoption is still modest.

A US Census Bureau Business Trends and Outlook Survey found only 20% of firms with more than 20 employees currently use AI in any business function. Among larger firms with at least 250 employees, adoption sits at 37%. Rajan himself calls that number "quite meager," given the survey's low bar for counting as an AI user.

Rajan is explicit that nobody knows the timeline. "AI-related job displacement is coming, though no one knows how fast it will proceed, how far it will go, and which sectors it will affect most," he wrote. He also cites companies "still running pilots and postponing hiring or firing decisions as they await more clarity."

This contrasts sharply with Sanders' framing of workers about to get "thrown out on the street."

The Case for Taking the Concern Seriously

Rajan is a serious economist, not a partisan actor, and his argument has a real mechanism behind it. If AI usage carries zero equivalent tax burden compared to a human employee, firms have an incentive to automate for tax reasons alone, separate from genuine productivity gains. That's a real distortion worth naming, even if it hasn't triggered mass layoffs yet.

Rajan also points to a genuine upside case. He argues AI could make remaining jobs more productive by removing drudgery, spur new categories of work like AI engineers who supervise the technology, and trigger what economists call the Jevons effect, where cheaper production from automation increases sales and, in turn, employment rather than shrinking it. He even floats AI lowering the cost of entrepreneurship, letting solo operators skip hiring a web developer or accountant.

The Free-Market Counterpoint

Richard Stern, an economist with the conservative think tank Advancing American Freedom, told Fox News this argument is really about two competing worldviews. "It's showcasing the difference between free market and central planning," Stern said, framing the debate as one about whether work is something people are "told to do" or something people choose to provide new value through.

Stern pointed to the Jacquard loom of the 1820s, an early punch-card weaving automation that threatened weavers at the time but didn't ultimately collapse employment in the sector.

Where Washington Actually Stands

The Trump administration's posture leans toward embracing AI rather than taxing it down. Breitbart News is hosting a policy event on August 25 in Washington, "AI Is Here: What to Fear, What to Embrace," featuring Emil Michael, the Under Secretary of War for Research and Engineering, along with executives from Substrate, ASI, and xAI. Taylor Budowich of Innovation Council Action, the event's sponsor, said Trump's AI agenda is "leading America into the future" of abundance.

That's the split. Sanders and House Democrats want a tax now, ahead of the disruption. Rajan wants a modest tax that scales with evidence as it comes in. Stern and the administration's allies want minimal intervention and faith in the market's track record of absorbing technological shifts.

The BLS numbers, as of the most recent data cited, still favor the wait-and-see camp. Whether that holds as AI adoption climbs past the current 20-37% range is the open question nobody, including Rajan, claims to have answered.