Research comparing labor force disruptions finds uniformly that automation effects are far larger than that of trade. This really shouldn’t be too surprising to anyone who cares to acquaint themselves with a few facts. Take steel, for example. Over the past 30 years, the production of American steel has risen about 10 percent, while employment has declined by almost 60 percent.

The steel example is helpful because it is so often mistakenly cited as an industry that needs trade protection. In fact, steel production would have to grow by 250 percent without any new technology to grow employment back to the levels of the late 1980s. That ain’t happening, folks, and we should probably not worry too much about it. Steel production is important, but steel jobs comprise an infinitesimal share of national employment. At levels of 35,000, there are more students at a typical Big Ten school than steel workers nationwide, and there are six times as many dog groomers nationally. 

Now, I’m not picking on steel workers; on the contrary, the marvelous gains in productivity are likely due in part to their skill and innovation. Productivity growth doesn’t come just from digital automation and robotics, but also from the insight and innovative ideas of plant employees. And that leads right to my next point about labor market disruption.

Productivity growth, whether through automation, plant design or better-skilled workers doesn’t kill jobs; it eliminates tasks. The first tasks we seek to eliminate are the hard, dirty and dangerous ones. Then we go after the routine non-cognitive tasks like those in an assembly plant or truck driver. That’s followed by eliminating the routine cognitive tasks like detecting cancer on a PET scan. 

Not all tasks have been eliminated, but the share of work done by manual human labor has been slashed profoundly over the past two centuries. I could fill several columns with the labor-saving technologies that freed us from raw manual labor, but suffice it to say few us carry our own water, cut and split wood for the stove, or hand loom our shirts. We do other things instead, and that is the crux of the issue. 

Automation and trade clearly disrupt many workers with redundant skills. My grandfather could easily harness a team of mules, my father calculated satellite orbits with a slide rule, and I am master of compass and sextant. My kids fancy themselves immune to technological redundancy because they are good with video game controllers. They should understand that Fortnite is merely the Frogger of 2018, and will look just as artless and technologically retrograde in only 40 years. 

As tasks are displaced by technology, demand for workers with those tasks will decline. We can predict which jobs are most at risk, but it is very challenging to time changes in demand. Fortunately, we also know which tasks are most likely to remain in high demand, for two reasons. 

First, we know the general trend of human consumption. The trend of consumption away from goods, towards services is a constant trend since the end of World War II. Services demand more labor, and increasingly more human capital-intensive labor. 

Second, we know the types of skills that are hard to automate. Artificial intelligence will allow the automation of anything that is routine, including tasks few now think will yield themselves to machines. Bricklaying, medical diagnostics, and anything involving the retrieval or processing of information, moving or tracking the movement of goods is at high risk. Truck drivers to economic forecasters will all have to find some new tasks to perform. 

Fortunately, we all possess skills that are hard to automate. The ability to perform non-routine tasks, purely human skills like empathy, the ability to integrate and analyze both quantitative and qualitative data at the same time, and most importantly the ability to learn new, non-routine tasks are the durable skills of the 21stcentury and beyond. 

These skills are taught at home and in schools. Indeed, the one place where all these come together are in middle and high school, and in traditional college settings. Programs that focus on career development and workforce training address little of these durable tasks. That should be a major policy worry across the United States, and especially in states such as Indiana with a large share of workers engaged in automatable tasks. It is these places that risk the greatest disruption from inevitable technological progress.