By: Julie Heath
Economists and policy makers are increasingly concerned about the hollowing out of the middle class—the computerization of jobs and entire occupations that once formed the backbone of a middle class lifestyle. Concern about computerization is not new. Many futurists have presented iconic images of robotic assembly lines for decades. In 1964, a group of scientists and economists—the Ad Hoc Committee on the Triple Revolution—warned President Johnson that computers would create massive unemployment, sooner rather than later. While the specter of mass unemployment did not materialize, technology has changed the labor market.
Computers can either substitute for, or complement labor. For example, in a Ford plant, robots have been programmed to apply adhesive that will hold a windshield in place and then install the windshield itself—both tasks that used to be performed by workers. On the other hand, computer-assisted surgery complements a surgeon’s skill, decreasing risk and improving patient outcomes.
The concern is that computers have developed to such an extent that they can increasingly substitute for labor in a variety of settings, not just on the assembly line. Bank tellers, airline ticket agents, utility meter readers, tax preparers, warehouse workers—all represent occupations that have been replaced, in whole or in part, by machines. And there are many more that are at the cusp of being replaced. The challenge, then, is to identify those skills that cannot be replaced by machines and ensure that our educational system is delivering them.
There are two characteristics of a task that, if met, mean that a machine can substitute for a human.
- All information necessary to perform and carry out the task is known and exists in a form that computers can process; and
- The processing of this information can be expressed in the form of rules.
Condition #1 may seem obvious, but it presents a significant stumbling block in what we would consider routine situations. As discussed in Dancing with Robots, authors Frank Levy and Richard Murnane provide an example of someone driving down the street, seeing a ball roll in front of the car. A computerized collision-avoidance system detects the distance the ball is from the car and “concludes” that this is not a dangerous situation. What the on-board computer does not know is that a rolling ball is often followed closely by a chasing child—a very dangerous situation. In other words, the computer does not have common sense. Using another of the authors’ examples, if you ask Siri the question: “Can a dog jump over a house?” she returns a listing of kennels, saying, “OK. One of these kennels looks fairly close to you.” Again, the computer fails on this task because a sufficient amount of common sense cannot (yet) be programmed into it.
The processing requirement is illustrated in the assembly line example. Computers can be programmed to move in physical space when the parameters are fixed. A continuous loop of information tells the computer that it is getting closer and closer to its target and it adjusts accordingly. A surgeon’s movements, on the other hand, are very hard to program since they can vary, depending upon the feedback he or she is getting from the computer and the particular patient’s needs. Likewise, moving furniture to a third floor apartment could not be performed by a computer.
In other words, computers can handle a vast array of tasks—of increasing complexity. But they are not flexible. They cannot anticipate problems, and they cannot respond in appropriate ways when situations arise that the rules-writers did not program. This can be clearly seen in the auto technician field. Even with the significant increase in computerization in this field, humans are still needed. Why? Because sometimes the computer cannot find the source of the problem and a human has to know where to start checking for the cause. The computer can follow the rules; a human has to be able to step in when the rules do not apply.
Since 1970, labor that is focused on routine cognitive tasks and routine manual tasks (assembly line work, for example) has decreased significantly. Non-routine manual tasks, solving unstructured problems, and working with new information have increased over the same period. This is a fundamental shift of the nature of work. Previously, a large number of tasks consisted of following directions. Now computers can be programmed to follow directions—workers need more fundamental knowledge.
What is that knowledge? Problem-solving. Critical thinking. The ability to communicate effectively. All things that we hope that our children are learning in school, but evidence indicates they are not. Or at least some of them are not. These new skills not only require the ability to read new information, but also the ability to put this new information in the proper context in order to work with it. Given the long-standing research on language development, children from more affluent homes have a decided edge in acquiring these skills because their parents talk to them more, have the means to engage them in enriching experiences, and have more books in their homes.
The National Governors Association spearheaded the development of the Common Core Standards in Math and English/Language Arts to place explicit emphasis on the critical thinking, problem-solving and communication skills needed in our increasingly robotized world. The fact of the matter is that we no longer live in a world where computers adapt to and complement human labor; human labor must find ways to adapt to and complement the tasks that computers can do. The implementation of the Common Core standards represents an acknowledgement of what those skills are and a pathway to transmitting them. It could also represent a way to slow the growing income inequality gap—if lower-income children can acquire the same amount and quality of robot-proof skills as more affluent children.