When Machines Take the Routine

Each new wave of automation brings predictions that machines will leave millions without work. History offers some reassurance: the mechanization of farming, for instance, eliminated most agricultural jobs in industrial countries, yet employment overall grew as new industries arose. Optimists conclude that technology destroys particular jobs but not work itself. That conclusion, however, glosses over an important detail. The new jobs created by past waves often required different skills, appeared in different places, and went to different people than the jobs that were lost. A middle-aged farmhand did not usually become a factory clerk; his children did. The transition, though beneficial in the long run, imposed heavy costs on a generation caught in between. The same pattern may be relevant to the Philippine business-process outsourcing industry, which employs large numbers of workers in tasks, such as answering routine inquiries and processing standard forms, that software is increasingly able to perform. The question is not simply whether new jobs will appear, but whether the workers displaced today can reach them. That depends on training that begins before layoffs, not after; on social protection that allows workers to retrain without falling into poverty; and on employers willing to redesign jobs around the tasks that people still do better than machines.