Robots and artificial intelligence are winning our jobs and an apocalypse is upon us, or so the employment horror story goes.
Innovation and automation are nothing new, nor is the fear they inspire. What’s changed is the interests of economists and policymakers who once tried to reassure anxious workers but now, after presiding over decades of labor market decline, are delighted to have a scapegoat. So we hear that the abandonment of the American workforce has been the unavoidable result of irresistible forces, the collateral damage of progress. And with those explanations come dire warnings for the future: Jobs will vanish, human workers will be rendered obsolete, reliance on a government check will become the norm.
It’s not true. Technology isn’t the culprit behind job loss, nor will it be. To the contrary, robots can be workers’ best friends. The abandonment of the American worker, instead, has been a conscious choice on the part of policymakers. And while that’s depressing, to be sure, it is also cause for hope. If bad choices are creating our employment challenges, better choices can solve them. The future of work is within our control, and technology is part of the solution, not the problem.
Getting comfortable with technology begins with understanding productivity, the measure of how much a worker can produce in some period of time. Imagine a widget factory in which 10 people each produce one widget each day. If the workers’ productivity doubles, each can make two widgets daily, and five workers can now make what once required 10. Did that improvement “destroy” five jobs?
Notice, I haven’t said anything about how productivity increased at the widget factory. Perhaps the owners installed a robot. That would be “automation.” But perhaps they switched to new materials that are easier to work with. Or provided training that improved everyone’s skills and thus their efficiency. Would anyone say that training “destroys” jobs or worry that it harms workers?
When people worry about automation, they are arbitrarily choosing one form of productivity growth to fear. But rising productivity is good for workers, regardless of its cause. Among other things, it is the necessary prerequisite to rising wages.
Hypothetically, automation could increase productivity so rapidly that many workers would became redundant faster than new roles for them could emerge. That would indeed be cause for concern. But it hasn’t happened in the past, and it is not happening today.
The government tracks worker productivity carefully and its data show that, however much technology is entering the workplace, productivity gains are actually slowing. In other words, technology is wiping out fewer jobs than ever. From 1950 to 2000, economy-wide productivity rose by an average of 2.1% each year, meaning that eight workers could produce by the end of each decade what required 10 to produce at the start. By comparison, productivity growth from 2000 to 2015 was 1.8%; and from 2010 to 2015, it was 0.7%. Far from living in an era of unprecedented job disruption, our era is one of relative stagnation.
Some studies predict that a large share of jobs will soon disappear, but those forecasts err by treating each job as a single task that either will or won’t be automated. When analysts look more closely at the many tasks that a given job entails, they find something more complex. Few jobs, perhaps 5% to 10%, appear fully automatable in the coming decades, with the technological advancements we can anticipate. For most jobs, some tasks can be automated but others can’t. The management consulting firm McKinsey and Co. estimates that roughly half of existing tasks could be automated by 2055. This mean we will still need human workers, and if we do things right, they will be much more productive.
Nor should we be worried that workers won't adapt. Business leaders constantly lament a “skills gap,” complaining that workers just don’t have the knowledge to implement new technologies. This is a market signal: Adopting technology in the workplace is going to be hard; it will proceed only as quickly as employers make it work with the employees they have. The path does not lead around workers but directly to them.
A brighter future does, however, require that policymakers grapple with what has actually gone wrong in our economy. We’ve built an education system that focuses on minting college graduates, though most Americans still don’t earn even a two-year degree. We’ve used environmental rules to make building things — widgets — in the real world too costly and risky. And our lax approach to international trade and immigration has encouraged businesses to look overseas for inexpensive labor, or bring it here, rather than find ways to partner effectively with the workers we have to make them more effective at their jobs.
If we instead make the American worker’s productivity our priority, there is every reason to be optimistic that our economy can once again be an engine of broad-based prosperity. We’ll be counting on technological progress to help.