It’s easy to look at the amazing advances in information technology and robotics over the last century and be fearful about the future of the American worker. From factory floors to your grocery store checkout, countless jobs once done by humans have been handed over to computers. Budding technologies like driverless cars promise that more of us will lose our jobs to a computer in the generation ahead.
But David Autor, a leading scholar of labor markets at M.I.T., offers a somewhat more sunny way of looking at things. In a paper presented at the annual gathering of central bankers in Jackson Hole, Wyo., on Friday, Mr. Autor argues that even as computers have gotten better at rote tasks, they have progressed far less in applying common sense.
Try to teach a computer how to tell that a picture of a chair is a chair, for example, and it will be befuddled. “Both a toilet and a traffic cone look somewhat like a chair,” Mr. Autor writes, “but a bit of reasoning about their shapes vis-à-vis the human anatomy suggests that a traffic cone is unlikely to make a comfortable seat. Drawing this inference, however, requires reasoning about what an object is ‘for,’ not simply what it looks like,” a skill computers generally still lack.
Machine learning, such as Google Translate or Netflix movie recommendations, is deeply inconsistent, he argues, “uncannily accurate at times, typically, only so-so; and occasionally, unfathomable.”
So what does that mean for workers over the years and decades ahead? Mr. Autor says that this weakness leaves plenty of opportunities for humans to serve as intermediaries of sorts between increasingly intelligent computers that nonetheless lack that common sense.
He invokes the idea of “Polanyi’s Paradox,” named for the Hungarian thinker Michael Polanyi, who observed that “we know more than we can tell,” meaning humans can do immensely complicated things like drive a car or tell one species of bird from another without fully understanding the technical details.
“Following Polanyi’s observation,” Mr. Autor writes, “the tasks that have proved most vexing to automate are those demanding flexibility, judgment, and common sense — skills that we understand only tacitly.”
So what does that mean for the jobs that will exist in the future, even as technology gets better and better at accomplishing many of the things that humans do now?
“Many of the middle-skill jobs that persist in the future will combine routine technical tasks with the set of non-routine tasks in which workers hold comparative advantage — interpersonal interaction, flexibility, adaptability and problem-solving,” Mr. Autor writes. He specifically mentions medical support jobs, building trades and some clerical jobs that require decision-making rather than typing and filing.
Continue reading the main story
14 hours ago
Boys and girls, please !!! This is more than machine intelligence. Robots or nearly-omnipotent machine entities of sci-fi like Colossus...
17 hours ago
Day tuk ur jeerrbs!Seriously though, Mr. Autor's rosy outlook is completely flawed, IMHO. Within 100 years, manual labor jobs will be...
17 hours ago
The problem is not technology. Technology has existed for all time. What is different is the economic system in which it operates....
See All Comments
Write a comment
In the paper, Mr. Autor presents data showing that these middle-skill jobs have indeed been under pressure over the last few decades, with much stronger growth in the number of both very basic low-paying jobs and the most advanced jobs for skilled professionals. It is a hollowing-out of the American work force, in effect, with fewer jobs for technicians and factory workers and the middle-class wages that come with them.
Continue reading the main story
But while acknowledging the trend in the past, Mr. Autor argues there’s not much reason to expect it to continue in the future.
“I expect that a significant stratum of middle-skill, non-college jobs combining specific vocational skills with foundational middle skills — literacy, numeracy, adaptability, problem-solving and common sense — will persist in the coming decades.” He argues that it is hard to blame computerization for jobs that have disappeared over the last decade in that much of the shift happened after capital investment in information technology fell following the collapse of the dot-com bubble.
Undergirding Mr. Autor’s optimism is the fact that mankind has consistently feared that technology will replace its jobs, and consistently been wrong. At the dawn of the 20th century, he notes, 41 percent of the American work force worked in agriculture, a number that fell to 2 percent by 2000. Farmers of that era could scarcely imagine that so few of their descendants would work in agriculture, or that so many would work in health care, finance, electronics, leisure and entertainment and so on.
“One can find fresh examples daily in which technology substitutes for human labor in an expanding — though still circumscribed — set of tasks,” Mr. Autor writes. “The complementarities are always harder to identify.” In other words, it is a lot easier to see the jobs that are endangered by emerging technologies than it is the opportunities for new jobs those technologies will create.