Top banner


Take That, Robots

To the Editors:

Brian Hayes’s great essay “Automation on the Job” (Computing Science, January–February 2009) resonated with my long involvement in automation debates and with my research and consulting work with robot assembly. In the 1970s, I was a member of the Automation Research Council that explored a wide range of technical, economic and social issues posed by the rise of Japanese manufacturing capability. We quickly learned that automation alone did not explain Japanese prowess. More important was a unique, close cooperation between labor and management as well as between design and manufacturing engineers.

Much of the worry about automation started with the assumption that robot labor would always outperform humans, which turned out not to be true. My graduate student Paul Lynch at Massachusetts Institute of Technology and I did the first economic analysis of robot assembly in 1976. We concluded that, except for special situations, assembly robots would have a difficult time competing with people economically or technically. Today, robots are used in niche applications where they are technically superior—for instance, in spot welding car bodies together, the first widespread use of robots, beginning in 1969, and still one of the largest.

Other situations suitable for deploying robots include those where local labor is expensive and scarce and products are carefully designed for automation (such as in Japan), where the work is dangerous (defusing bombs), where parts to be assembled are very light and simple, and where products are basically flat (components on circuit boards, electronic watches and cell phones, for instance). Many small, light and simple things produced in huge quantities (cigarettes, spray bottle tops, felt tip pens, razor blade cartridges) are still made by special-purpose machines, not robots. If the parts weigh more than about 100 grams and the product is three-dimensional, almost all assembly is manual.

Many U.S. jobs that people believe were lost to automation were actually lost to low-cost manual labor in developing countries.

Daniel E. Whitney
Arlington, MA



Bottom Banner