The GOP Wants to Put Workers Under AI's Thumb: A Shorter Work Week Is a Better Answer

Productivity growth is an old concept; we’ve been seeing it at a substantial pace for more than 200 years. Nonetheless, many elite intellectual types like to claim they know nothing about it when they talk about AI.

It’s far from clear how much of a productivity boom we will see with AI. For people who are lost with my reference to productivity growth, the story that AI will take all the jobs is a story of a massive productivity boom. If that happens, it will mean that the people who are still working will be hugely more productive, since we will be producing the same or more goods and services as we do at present, with many fewer people working.

FWIW, virtually no major forecaster or forecasting agency is projecting anything like this productivity boom. For example, the Congressional Budget Office (CBO) projects that productivity growth will average 1.5 percent over the next decade.

That’s a healthy rate of productivity growth, but nothing extraordinary. It’s a bit better than the 1.3 percent rate from 2005 to 2025, but less than the 2.0 percent rate we saw in the 1990s and much less than the 2.4 percent pace the country had from 1947 to 1973. There is no story of AI creating mass unemployment here.

CBO is not God, but they are pretty much in the center of professional forecasters by design. They try to make sure that their forecasts do not vary hugely from what other public and private sector forecasters are projecting.

It is also worth noting that if CBO is seriously wrong on the low side, then some other things logically follow. Most importantly, if productivity growth proves to be far more rapid than what they have projected, GDP growth will also be far more rapid than projected. This would mean, among other things, that the debt-to-GDP ratios will be much lower in the future than is currently projected.

In other words, the people yelling about unsustainable debts and deficits need to STFU. You can’t both be expecting a massive AI productivity boom and think the US has a huge debt problem. That is not a matter of opinion; it is a matter of logic.

But let’s assume for a moment that we do get a huge productivity boom from AI. We don’t need to run around like chickens with our heads cut off when we ask what to do about it. Because productivity growth is in fact a very very old phenomenon. We have long known how to deal with it; we shorten work hours.

Workers in Germany, France, and other wealthy countries work on average 20-25 percent fewer hours a year than Americans.

That is why we got the 40-hour work week with the Fair Labor Standards Act (FLSA) in 1937. The Act doesn’t actually prohibit employers from having longer work weeks; it simply requires them to pay a 50 percent premium for overtime hours. This was supposed to encourage them to hire more workers instead of working their existing workforce more hours. (Contrary to the way it is discussed in the media, the decision to put in overtime is almost always the employer’s, not the worker’s. Unless a union contract specifies otherwise, an employer has the option to fire a worker who refuses overtime.)

This is why it was truly incredible that Trump eliminated the income tax on the overtime premium. This is effectively encouraging employers to have longer workweeks, 180 degrees opposite of the intention of the FLSA.

But just because Donald Trump and Republicans in Congress have no grasp of economics doesn’t mean the rest of us shouldn’t. If we really are seeing an AI-driven productivity boom, the most obvious way to deal with it is to shorten the workweek and work year. The United States is an outlier here. While we were originally a leader in implementing a 40-hour workweek, we have done little to reduce work time in the 90 years since then.

As a result, workers in Germany, France, and other wealthy countries work on average 20-25 percent fewer hours a year than Americans. As a crude approximation, if workers put in 20 percent fewer hours on average, it will mean 20 percent more jobs. Things in the real world are never quite that simple, but the basic logic that shorter work years means more jobs does hold.

It’s also not rocket science to get to shorter work years. We can amend the FLSA so that the overtime wage premium kicks in at 34 or 36 hours. Also, instead of removing taxes on the premium (having taxpayers subsidize long workweeks), we can raise the premium from 50 percent to 75 percent, as the Congressional Progressive Caucus recently proposed. We can also mandate 2 weeks or more vacation, along with paid sick days and family leave, as many states have already done. All this is old-fashioned stuff that other wealthy countries have been doing for decades, and we have done in the United States nationally in the distant past and more recently at the state level.

The immediate prompt for this diatribe was a New York Times article that asked how we will deal with a collapse of employment from AI. In fairness, the piece does note that an AI-driven productivity boom is far from certain, but it then suggests that if it does happen, a universal basic income, or a universal high income might be ways to deal with it. The piece notes that Elon Musk is supposedly an advocate of the latter.

While any pro-worker legislation will face an enormous uphill battle in the current political environment, a variation of policies that people have seen for a century might have a better shot than something that seems completely new.

While these proposals are, in principle, fine, they ignore the reality of US politics. Just four years ago, when the Democrats had a trifecta, they could not get a modest increase in the child tax credit approved in the Senate. Get out your yard stick and try to measure the distance between a modest boost to the child tax credit and a universal basic income, much less a universal high-income.

It’s probably also worth mentioning that Elon Musk has done everything he can to keep his workers at Tesla from forming a union, where they would be better able to secure their share of the company’s profits. That may lead reasonable people to question his commitment to workers’ well-being in an era of AI-driven mass unemployment.

While any pro-worker legislation will face an enormous uphill battle in the current political environment, a variation of policies that people have seen for a century might have a better shot than something that seems completely new. It is also worth pointing out that the tools for dealing with a surge in productivity growth are well-known and tested. Whether or not a universal basic income is a better way to go, our toolbox is already far from empty when it comes to dealing with this situation. This is not a new story, and it is wrong to portray it as one.