April 10, 2019


Are You Sure You Need a Human?: A Form From the Future looks at what humans and machines do best. (ROSE EVELETH, APRIL 10, 2019, Slate)

Let's say you're a manager at a big box store or a fast-food franchise or even a hospital. You might also need to fill out paperwork justifying your request for an actual person to pull boxes from a shelf, check people out at a store, deliver medication, or make a hamburger. How long do you think you'll need this human for, and who is responsible for her safety? Do you really need a human? Or could a machine do the same job for longer and for cheaper?

Trying to quantify the projected length of a person's working life can be traced back to the early 1900s in the United States. Nineteenth-century life insurance companies relied on mortality tables provided by cities and churches to calculate their policies and payouts. In the early 1900s, AT&T became one of the first huge companies to adopt the idea of a "service life" for its machinery--keeping track of things like telephone poles and vehicles and gear and trying to predict when they might fail. Today, the concept of a "service life" for a piece of equipment is commonplace, but at the time it was relatively unusual. "These graphs allowed them to predict the lives of their machinery," says Dan Bouk, a historian and the author of How Our Days Became Numbered. They meant that the company could replace things before they were run into the ground and failed. Then, in 1927, the opportunity arose to apply the same thinking to people.

The late 1920s saw the rise of pensions beyond the military in the United States. Which meant that huge companies like AT&T suddenly had an incentive to predict how much work they were getting out of their employees and how much they might be expected to pay them back once they retired. Just like with their telephone poles, vehicles, and gear, AT&T began collecting data on everything from employment tenure to wages. The company designed tables to better understand its workforce, "which they can use to predict what the service life of their various employees will be," Bouk says. With this data, it was able to offer retirement to workers who might not have been thinking about it but who the company predicted would no longer provide enough work or efficiency to be worth employing.

Today, the conflation of service life and human life is less about pensions and more about mitigating other kinds of financial risks. Gone are the days when many workers spent their entire professional career at a single company. In a world where Amazon has patented (although doesn't seem to have actually developed) a cage to protect humans from their robotic co-workers, companies looking to get more work for less money (and trouble) are turning to machines that won't complain or get sick or unionize. Machines usually have a higher upfront cost, but they don't require health care or safe working conditions. 

Posted by at April 10, 2019 4:19 AM