Why Leaders Need To Know What Machines Can't Do

Just because a computerized device can do a job does not mean it will perform that task better than a human. JL

Geoff Colvin reports in Fortune:

As machines grow more powerful, deciding who must go and who must stay becomes harder. A guiding principle: Just because technology can do a job brilliantly doesn’t mean that it should.
When stock markets plunged early this year, managers at USAA’s investments division noticed something odd. Customers who routinely conducted business online were suddenly lighting up the phones. USAA had nothing new to tell them—its fundamental advice hadn’t changed, and they could have found that guidance online. Yet clients deeply wanted to talk to a real human being, and never mind why. They just did.
That reality illustrates a high-stakes decision that confronts managers in every industry: choosing which employees must be replaced by technology and which must not be. Growing numbers of jobs at every level can be performed by ­machines—not just faster and more cheaply than humans can do them, but better. In many of those jobs, such as in factories, failing to replace people could doom a company through uncompetitive costs. Yet in other jobs that machines can do well, such as giving financial advice, replacing too many humans could be a fatal error. How to decide? Three situations in particular seem to justify the costs, and quirks, of people.
When customers value the human touch. Many decisions that in theory are calculable—where to invest, whether to sue, how to respond to a medical ­diagnosis—are in fact laden with emotion. Many people need to interact with a person before choosing a course of action. In finance, law, medicine, and other fields, workers who handle those interactions most adeptly will be the least susceptible to replacement.
When constituencies must be represented. All organizations are run ultimately by and for humans, and most are complex matrices of desires, incentives, budgets, and myriad other factors. If marketing can’t get along with sales, or management with labor, nothing good can happen. Technology could optimize the whole intricate machine, but it will seize up if humans can’t agree on how to make it go.
When someone must be accountable. So long as humans and not machines are in charge—let’s assume that’s a long time—societies will demand that people be made to answer for decisions, even if technology recommends those decisions. Government officials, military officers, judges, business managers, basketball coaches, and others in leadership roles will remain where the buck stops. Technology may reduce the number of people in such roles—it’s already taking over tasks of middle managers, for example—but responsibility will ultimately end up in human hands.
As machines grow more powerful, deciding who must go and who must stay becomes harder. A guiding principle: Just because technology can do a job brilliantly doesn’t mean that it should.