Here’s the dirty little secret about automation: it’s easier to build a robot to replace a junior attorney than to replace a journeyman electrician. And that fact helps explain why economists and politicians are feeling misgivings about “creative destruction,” which, up to now, they have usually embraced as a net good for society. Technology and automation, they’ve argued—correctly—boost productivity and create more jobs overall (even as some kinds of work get eradicated).
In the age of the algorithm, though, they’re not so sure any more, and no wonder: instead of creative destruction coming to factories and farms, it’s sweeping through city centers and taking white-collar jobs. The chattering classes have talked and written for years about the “end of work.” Doubtless many fear that the end of their work is in the offing, this time around.
Understandably, most of the media focus has been on the replacement of manual labor—real robots doing real tasks provide better visuals than an invisible service “bot” in the cloud. But focusing on hamburger-flipping androids is a distraction from where the real revolution is taking place. In any case, automation doesn’t really explain the decline in factory employment: manufacturing-sector investment in information technology has been flat or declining for more than a decade; and productivity, a critical indicator of (and the purpose of) automation has also been flat. Factories are actually underinvested in technology.
But Silicon Valley has been busy building software that will transform businesses, particularly those associated with shopping malls, Hollywood, hotels, taxis, newspapers, broadcast TV, finance, and even education. Some algorithms can teach basic math better than most humans with a B.Ed. Next up are the many paperwork tasks currently administered by bureaucrats and regulators. The white-collar professionals who staff these service-sector domains, a vast cadre, may find themselves displaced.
Consider a bellwether of more white-collar disruption yet to come: of the nearly 200 so-called Unicorns—private, venture-backed companies such as Uber that are valued over $1 billion—90 percent are in nonmanufacturing businesses. There’s a good reason for such a skewed focus. Supercomputer-class software in the cloud can perform, at minimal cost, once-daunting information-centric tasks, from reading X-rays to managing a “passive” investment fund. But the engineering challenges are far greater and many times more complex in cyber-physical systems, where software meets steel in real time. A seemingly minor software glitch that freezes a video screen can escalate from an annoyance to a fatal error if it’s involved in the interaction of the mechanical world with human beings. Self-driving car hype aside, there’s much work to be done in achieving viable sensors, actuators, power systems, security, and safety. Goldman Sachs reports that the automobile sector remains the only deeply automated industry; nearly all the rest are, at most, 20 percent to 30 percent along this path.
In due course, a cyber-revolution will indeed come to the “means of production.” Meantime, we’re in the midst of an upheaval in what we might call the “means of management.” The overall effect, I believe, will be the same as in the past—a boost to the economy and more jobs—but the makeover this time will affect the professional and managerial classes. We should expect them to be at least as vocal about it as many factory workers were a generation ago.
Photo by Visual China Group/Getty Images