One of humanity’s oldest pursuits is inventing machines that reduce the labor-hours needed to perform tasks. History offers hundreds of examples, each of which seemed amazing at the time. In the modern age, we have seen the arrival of the automatic washing machine in the 1920s; the programmable logic controller (PLC), which enabled the first era of manufacturing automation in 1968; the word processor in 1976; and the first computer spreadsheet in 1979. And none of these or similar inventions triggered an era of endemic unemployment.
But now that America is finally approaching full employment, a chorus of experts claims that this happy situation will be short-lived. Their warning is not the usual one about unemployment rising again when the next recession hits. Instead, the proposition is more ominous: Technology is now finally able to replace people in most jobs. It is the age-old mantra of failed forecasters: This time it’s different.
Consider one of the most powerful advances in the history of labor productivity: the arrival in 1913 of a practical automobile. The very word “automobile” was coined to describe the automation of mobility, a long-time goal of humanity. The transition from grain-fed horses to petroleum-fueled automobiles took place with astonishing velocity precisely because the productivity benefits were so profound. Along the way, the automobile totally upended the character and locus of every kind of employment associated with transportation. Gone forever were all the jobs and the millions of acres of land devoted to the feeding, care, and use of horses.
Here’s another, more recent, example. In 1956 the invention of the shipping container also radically increased labor productivity. Containerization led to a 2,000 percent gain in shipping productivity in just five years and the end of centuries of rising employment for longshoremen (and, historically, they were all men). The gain in productivity rivaled that which had taken place in the previous century with the switch from sail to steam power. Both innovations radically lowered shipping costs, and both propelled world trade, prosperity, and employment.
Some jobs were lost of course. As with the “typing pools,” rows of accountants “ciphering,” and rooms full of draftsmen with sharp pencils — among other common workplace sights from bygone eras — it is easy to predict which groups will lose jobs from technology-driven gains in labor productivity. It is far harder, on the other hand, to predict what kinds of newjobs will appear, and how many.
It’s also easy to see how an economy expanding due to accelerated productivity will produce added wealth, which, in turn, leads to more demand for existing products and services, all of which entail labor. It is far more challenging, however, to predict exactly how new technologies will enable unanticipated businesses to emerge to produce new kinds of products and services, all employing people in entirely novel ways.
But will it be different this time? For the first time in history, will the burgeoning expansion of an entirely new technological infrastructure — AI and robots — finally destroy more jobs than it creates?
We know two things about the effect of the myriad technology changes over the past 130 years. The first is that continual and often profound advances in labor-saving productivity have boosted the economy so much that per capita wealth has reached unprecedented heights. The second is that despite all that “labor saving,” about 95 percent of willing and able people have, on average, continued to be employed. In other words, the unemployment rate has remained essentially unchanged, at about 5 percent, for all of those 130 years, fluctuating episodically due to cyclical recessions.
If labor-saving technology were inherently a net job destroyer, the unemployment rate should have risen inexorably throughout history. But it hasn’t. MIT economist David Autor is particularly eloquent on the apparent paradox of more employment despite inexorable advances in labor-reducing technologies, observing that, with regard to the prospects for employment growth, “the fundamental threat is not technology per se but misgovernance.”
The truth is that we have yet to experience the full benefits of AI and robotics. Nearly everything that is being claimed about labor destruction is based on speculation and flawed forecasts. It is still early days. It took time to conquer the complexities of manufacturing reliable, affordable cars and to develop all the associated industries and infrastructures. But once that happened, things really took off.
In the first two decades of the automobile, very few people owned cars. By 1937, the percentage of car ownership reached 25 percent; by 1957, it was 40 percent; today, it stands at 85 percent. Along the way, the entire structure of the American economy was transformed, stimulating growth and new employment across the landscape. Detroit was the Silicon Valley of its day, and the benefits of ubiquitous personal transportation boosted the entire nation.
Every major infrastructure shift in U.S. history has propelled the nation to greater prosperity. Every era has created more new companies and jobs — most of them entirely unanticipated — than were lost from the former, less productive centers of the economy. The central challenge of our era is not the prospect of the end of work and systemic high unemployment. Our challenge is the same as that of earlier eras: We have a moral and political imperative to deal with the inevitable disruptions that occur along the way to enhanced prosperity.
When it comes to practical AI and useful robots, we are today — to put it in terms of the automobile age — still circa 1920.
Mark P. Mills is a senior fellow at the Manhattan Institute and a McCormick School of Engineering Faculty Fellow at Northwestern University. This article is based on Mills’s book, “Work in the Age of Robots,” published this month by Encounter Books.