Ahead of a hearing today for the Senate Health, Education, Labor, and Pensions Committee, Senator Bernie Sanders released a new report with a startling title: “The Big Tech Oligarchs’ War Against Workers: AI and Automation Could Destroy Nearly 100 Million U.S Jobs in a Decade.” The report warns that “a new technological age stands to deepen this war against workers and increase economic inequality,” arguing that AI-driven job displacement is a crisis that requires policy intervention. Among a variety of other solutions, the report calls for robot taxes on large corporations that will be used to benefit workers displaced by AI.

While I commend Senator Sanders’ staff for its pioneering use of ChatGPT to model job losses, there are a number of serious issues with the report and its proposal to tax robots.

For one, the report reviews some key papers on automation and income inequality, but nowhere does it review the current literature showing that new AI tools are reducing inequality. In Brynjolfsson et al. (2023); Caplin et al. (2024); Choi et al. (2023); Hoffmann et al. (2024); Noy & Zhang (2023); and Hauser & Doshi (2024), advanced AI tools were found to be skill equalizers, raising the performance of those at the bottom in customer support, legal work, and software development, among others. If Sanders was truly concerned with worker inequality, he should be optimistic about  AI tools and engaging with the empirical work on this subject, which I have been actively collecting here.

Second, the report doesn’t contextualize what 100 million jobs lost over a decade would mean for an economy of 170 million workers. In any given month, roughly 4 to 6 million jobs are lost. Over a decade then, the US economy experiences roughly 480-720 million job separations through the normal churn of business, putting the 100 million job loss number at or below 20 percent of total job turnover that would occur regardless.

Finally, modeling out AI job loss is a tricky business. In an extended review of AI job loss predictions back in 2019, I pointed out that a small shift in methodology between estimates from the University of Oxford and from PwC had the effect of completely changing the impact of AI, causing job loss estimates to swing from 47 percent to 9 percent. And much like those earlier estimates, Sanders’ model says little about the cost-benefit tradeoffs that firms face when adopting new tech or how automation will interact with task reallocation to determine net employment effects.

Should Robots be Taxed?

Still, Sanders is not alone in calling for a robot tax. Bill Gates, Mark Cuban, Bill de Blasio, and Jeremy Corbyn have all called for taxes on robots. These proposals tend to come up short because of two problems: one as a result of defining robots and automation and the other as a result of designing an efficient tax.

Advocates for a robot tax tend to sidestep the thorny issue of writing a bill because it is incredibly difficult to formally define robots. For example, Automation Direct, a major distributor of industrial robotics equipment, has many products that come in both manual and automated variants, such as  the programmable logic controllers (PLCs). Low skilled labor often uses PLCs in semi-automated production lines where workers still handle machine operation and then conduct quality checks between steps. Imposing a significant tax on PLCs wouldn’t just affect manufacturing jobs, it would force companies to substitute to less automated computing systems to achieve the same results.

Overly broad definitions have plagued AI bills as well. As I explained in a recent post, California’s AB 1018 was written to bring back a human touch to automated decision systems in housing, health, and finance. Thankfully, the bill failed to pass the legislature. It was so broadly written that even Excel spreadsheets would have triggered regulatory requirements.  

Then, there is the problem of economically efficient taxation.

Finding the efficient tax rate comes after determining how large a program should be. It is left up to the reader to imagine how big an AI adjustment program will need to be. Still, it isn’t as though companies aren’t already being taxed. Capital gains taxes already apply to productive AI investment; corporate income taxes already capture profits from automation; and property taxes often include machinery and equipment. Stacking a robot tax on top of these would be redundant.

Critics might say that this time is different. But even if they are correct, this is an argument for better transition assistance, not just a tax on one arbitrarily defined category of capital goods.

Senator Sanders’ report raises important questions about worker welfare in an AI world, but in many ways, the report comes up short.