For the last few decades, we have struggled with how to employ manufacturing workers who lost their well-paid job with great benefits due to a globalized economy. When workers in another part of the world are willing to work for a fraction of what it costs to manufacture something in the United States, it’s obvious why companies move their manufacturing operations: it’s a significant cost advantage and, worse, if they don’t, their competitors will. This is only more true today. In January, Charles Duhigg and Keith Bradsher reported for the New York Times that for technology products especially, the labor cost itself is less important. What matters is that Asia—especially China—is the only place where every part of the supply chain exists in one region, that can manufacture quickly and at immense scale.
Manufacturing, too, is increasingly automated. The human’s role in actually putting things together is decreasing. Automation on large scale for identical products, like cars, has been a reality for decades. What’s happening now, though, is that smaller scale, small production runs are being automated as well. Rethink Robotics has created a robot called Baxter that can be “taught” how to do repeating tasks, and can work around humans. Rethink Robotics says Baxter can work for the equivalent of $4 an hour. Vanguard Plastics, a 30-person company in Connecticut, is using Baxter for menial tasks. Vanguard’s president, Chris Budnick, says that workers who did these jobs before are not being laid off, but are now assigned to “higher-level” tasks like training Baxter for each new production run.
Robots like Baxter are a work multiplier. Whereas before Vanguard required humans to do menial tasks, now they only need humans to train robots how to do something. But many more people are required to do the menial tasks than are required to train robots, so while no one may be losing their job now, they will need to find new productive tasks for them in the future—or eliminate their jobs. As robots like Baxter get better, too, manufactures will need even fewer employees to train them.
Other industries face very similar problems. Retail salespersons and cashiers, for example, account for nearly 6 percent of all jobs in the U.S., but are increasingly irrelevant. For many products, shopping online is more convenient and cheaper. Tower Records, Blockbuster and Borders all failed fundamentally because purchasing music, movies and books online is much better than paying more money for the privilege of driving to a store, hoping they have what you want and waiting in line. Even grocery stores are reducing their need for cashiers by employing self-checkout machines, which allow customers to scan and pay for items on their own and require only one employee to monitor several self-checkout machines.
Almost all of the jobs lost due to offshoring and automation have been low or semi-skilled kinds of jobs. Manufacturing jobs required training, but certainly did not require several years of specialty education and training to do. Retail sales and cashier positions require almost zero training. It would appear, then, that since offshoring and automation are eliminating low and semi-skilled jobs, we can re-orient our economy toward “knowledge work,” or work whose primary task is thinking. Examples of these kinds of jobs are software engineers, engineers, lawyers, doctors, accountants, managers and scientists. These kinds of jobs require a tremendous investment in education and training, and therefore seem not to fall prey to offshoring and automation.
In The Lights in the Tunnel, Martin Ford asks a very good question: “What is the likely economic impact of machines or computers that begin to catch up with—and maybe even surpass—the average person’s capability to do a typical job?” Or, more provocatively: If computers can already beat the best chess players in the world, isn’t it likely that they will also soon be able to perform many routine jobs?
Ford argues that not only is this true, as we’re seeing for manufacturing and retail jobs, but that it is also true for highly-skilled knowledge work jobs. Think about what a radiologist does. Much of what they do is read routine x-rays or CT and MRI scans to diagnose issues with patients. Since radiology is increasingly digital, and knowledge of what different conditions and diseases look like can be digitally represented and algorithmically identified, it’s likely that some of what human radiologists do today—the more routine, easy to identify cases—will be handled by computers instead. Doing so will dramatically decrease costs for hospitals because they will have to employ less doctors, which require large salaries, health insurance, vacation and sick days, and have to be hired and managed. Computers don’t.
The same, of course, is true for much of what general practice doctors do as well. Computers like IBM’s Watson could diagnose patients with routine things like the flu and provide a treatment as well. In fact, because Watson would have access to exponentially more medical research, journal articles, studies and patient history (and aggregate patient data), Watson may very well provide better diagnoses and treatments than the average human doctor.
Ford points out this is true for other fields, too, like law. He writes:
Currently there are jobs in the United States for many thousands of lawyers who rarely, if ever, go into a courtroom. These attorneys are employed in the areas of legal research and contracts. They work at law firms and spend much of their time in the library or accessing legal databases through their computers. They research case law, and write briefs which summarize relevant court cases and legal strategies from the past.… Can a computer do the lawyer’s job? (70-71)
Is there any reason to think that computers will never be able to do this kind of basic research and summarization? I don’t think so. What this suggests is that automation will challenge many kinds of knowledge work just as much as low and semi-skilled work. Indeed, companies will have even more reason to automate these kinds of jobs, because they are generally very well-paid jobs.
Manufacturing and retail job elimination, then, is just the first wave of many to come. The question, though, is not how to get those jobs back and protect the ones that still exist. That isn’t going to happen, is counter-productive and a waste of time. The question to ask is, when many of the jobs people depend on our automated, what kind of jobs will they do instead?
That question is, I think, the most important question to answer for the next few decades.
I have some ideas, but for now, I just want to ask the question and want you to think about it. How do we productively employ these people?
