Levy and Murnane contend that computer technology replaces workers if their jobs can be reduced to steps based on rules, no matter how complicated the steps or rules, because computers are very good at rule-based steps. This category includes both low-wage jobs, such as bank teller, cashier and telephone operator, and high-wage jobs, such as loan officer, market floor trader and welder. Computers can even program themselves, replacing at least some computer programmers, if the programming steps can be defined well enough.
Computers do not replace all low-wage jobs because some jobs cannot be reduced to rules, such as driving a truck, cleaning a room or caring for children. These jobs require judgment based on information about many variables in real time.
Computers also cannot effectively negotiate conflicts and cooperation between people, because negotiation requires recognition of fine lines in arguments and adjustment to subtle signals between negotiators. Making a left-hand turn, for example, requires constant real-time negotiation with other drivers and pedestrians. Computers enhance rather than replace jobs that require judgment: doctors evaluating patients, financial analysts deciding the worth of a merger and auto mechanics troubleshooting a car. Computers also enhance rather than replace jobs that require negotiation: lawyers representing clients and teachers instructing students.
The authors, professors of economics at MIT and education at Harvard, respectively, break down intelligence into finer components in order to understand the division of labor between rules-based and judgment-based tasks. Intelligence combines information; pattern recognition; and structural schemas that recognize concepts, organize them into patterns and connect them with other concepts. Computers can do this in simple ways; they can recognize a request from a customer at an ATM, retrieve account information and perform the instructed tasks. But computers are not imaginative enough to think—at least not yet.
Computers also cannot facilitate subtle interactions between workers. Sharing information among colleagues, for example, requires cooperative skills, retention and communication.
Rules-based jobs pay little because those who do them can always be replaced by computers. But why do some judgment-based jobs, like medicine, pay well, and others, like minding children, do not? The answer is that those people replaced by computers compete for jobs in the service sector that cannot be replaced by computers but that are not computer enhanced, either.
This is not the end of the story, of course, because many jobs replaced by computers, and even those enhanced by computers, can easily move away. Computers can communicate quickly and cheaply over great distances using the least expensive information technology. Neither side of the New Division of Labor favors union organizing: Jobs that use technology have been difficult to organize, and unions have not gained much ground in low-wage service jobs.
For teachers and professors, Levy and Murnane’s lesson is simple: Keep teaching students to think and work together. College graduates are paid more than high school graduates not simply because they can use computers, which admittedly increases their value in the job market, but because they can make complex judgments, negotiate and share information—the core of the liberal arts.
Dan Georgianna is a professor of economics and president of the University of Massachusetts-Dartmouth Faculty Federation. This article is a product of the Future of Work Project, sponsored by the Massachusetts AFL-CIO and the labor education centers at UMass Amherst, Boston, Dartmouth and Lowell.











