Don’t tax the robot — enjoy the windfall

Back in 2017, South Korea quietly cut tax credits for firms investing in productivity enhancing equipment. Large companies saw their credit shrink from 3 percent to 1 percent and mid-sized firms from 5 percent to 3 percent, while only small firms kept a 7 percent benefit. To some, it looked like the beginnings of a “robot tax.”

Eight years later, that debate is back — and louder.

With fears of AI (artificial intelligence)-driven disruption in the news almost daily, proposals for a robot tax or even a “token tax” are now being floated in Korea, the United States and beyond. The argument is as simple as it is wrong: As jobs are lost, aggregate demand shrinks, and therefore governments should tax the firms driving automation and use the proceeds to fund income support.

Last fall, then-opposition leader Lee Jae Myung, now Korea’s president, declared that if the public loses jobs, “purchasing power must be preserved.” His solution: Levy a charge on AI and robots to fund economic sustainability. Anthropic CEO Dario Amodei went further, proposing that governments collect a 3 percent token tax on AI revenues to redistribute the gains.

These proposals are gaining traction in Seoul. They echo a broader anxiety worldwide that somehow this time is different and unlike past major technology innovations, AI will somehow lead to permanent higher unemployment. But while the impulse is understandable, the theory — if it can even be called that — is completely wrong. Taxing robots or tokens is the path to stagnation.

Why robot and token taxes are misguided

First, they hit productivity exactly when economies need it most. Korea is aging faster than almost any nation on earth. Its working-age population is shrinking. If Korean firms slow their adoption of automation because of new taxes, the country will see weaker productivity, slower growth and fewer resources to support its aging population.

Second, they are based on panic. It’s much more likely that AI is a “normal” technology, just as the steam engine and electricity were, boosting productivity from maybe 2 percent a year to 4 percent. It’s way too early to panic about AI job loss when Korea doesn’t even include AI in its productivity statistics. Take a deep breath and calm down.

Third, they rest on a political illusion. Taxing robots doesn’t magically save jobs. It discourages firms from investing, which in turn erodes competitiveness. Over time, both wages and employment decline. We have seen this before — in the industrial revolution, in agriculture, in IT. The solution was never to tax the tool. It was to help workers move into new industries.

Fourth, they are based on what economists call the “lump of labor” fallacy: the idea that there is a limited amount of work to be done, and if a job is eliminated with technology, it’s gone for good. But this is a false reading of the process of technological change because it fails to include second-order effects whereby the savings from increased productivity are recycled into the economy in the form of higher wages, higher profits and reduced prices. This creates new demand that in turn creates other jobs, some in new occupations (like “content creator assistant”) and many in existing ones that workers will now spend more money on (e.g., personal trainer). A long tradition of economic analysis shows zero relationship between productivity growth and overall unemployment growth. Zero.

Boost AI adoption and help workers make transitions

Rejecting robot or token taxes doesn’t mean ignoring the problem.

First, don’t assume that Korea will benefit from fast productivity growth from AI. It’s just as likely that it won’t. So restore the investment tax credit to 10 percent for all general firms, big and small, and 25 percent of national strategic technologies.

Second, instead of using robot taxes, recycle the productivity gains and the taxes they generate into more funding for education, retraining and labor mobility.

Korea’s teachable moment

The current debate in Korea shows the risks of getting this wrong. Business leaders warn that robot or token taxes will slow investment just when they need to scale AI across logistics, retail and manufacturing. Academics note that delaying automation will not protect jobs, but will erode the competitiveness that sustains them. The Bank of Korea has cautioned that AI adoption reduces labor demand even as it raises skill requirements, and that Korea’s rigid dual labor market makes transitions especially hard.

Incentives matter. If governments tax robots, firms will buy fewer of them. If governments instead reward companies that adopt AI while maintaining or even growing employment, then technology and jobs can advance together. Carrots work better than sticks.

In other words, don’t tax the robot — help workers adjust.

Dr. Robert D. Atkinson (@RobAtkinsonITIF) is president of the Information Technology and Innovation Foundation, an independent, nonpartisan research and educational institute focusing on the intersection of technological innovation and public policy. Kim Se-jin is a tech policy analyst for ITIF's Center for Korean Innovation and Competitiveness. The views expressed in the above article are those of the authors.