When most people think about intelligent robots, their images are typically of mechanical simulations of humans, epitomized perhaps by the character C3PO from the Star Wars films. InFlesh and Machines: How Robots Will Change Us, Rodney A. Brooks reverses that view, showing how humans are essentially machines. He also shows how he has achieved success by negating the basic assumptions of other researchers in the artificial- intelligence community.
Early goals in artificial intelligence were mostly “things that highly educated male scientists found challenging,” such as chess, integral calculus, and complex algebra problems. Actions that small children can do effortlessly, such as visually identifying a table or cup or walking from bedroom to living room, were not considered to require intelligence. Walking on two legs and seeing, however, turned out to be harder to accomplish than the loftier-sounding early goals.
In the mid-1980’s, Brooks looked anew at the foundation of research up to that time, in which a robot’s sensors and motor processes were coordinated by a “cognition box.” The best way to build this center of thinking and intelligence, he decided, was to eliminate it. While others worked on how best to have a robot map the world around it and then use that map to move about, Brooks took the contrarian angle and eliminated the map, so his robot would simply sense and use the immediate environment. This is the way real-world animals operate, he reasoned.
This is also Brooks’s major approach to problem-solving. Early in his career, he “would look at how everyone else was tackling a certain problem and find the core central thing that they all agreed on so much that they never even talked about it.” Then he would throw that thing out and see where it led. For example, when everyone else was working on how to represent certain obstacles, he looked at “where the stuff wasn’t.”
By looking at the evolution of real insects and other animals, he noted that they started with simple capabilities and gradually developed more sophisticated capabilities. The way to build robots, he reasoned, was to build complex capabilities on top of simpler ones. An early exemplar of this approach was Genghis, an insectlike creature that detected the warmth of any passing person, then used six insectlike legs to walk over anything in its path as it followed the person around like a puppy. It had no large master control program, but a set of fifty-one relatively small programs, each of which sensed a certain thing or actuated a certain motion. In Flesh and Machines, Brooks devotes a twelve-page appendix to describing the design and implementation of Genghis.
Raised in an isolated part of Australia “at the end of the technological Earth,” Brooks says he “grew up a nerd in a place that did not know what a nerd was.” By age twelve, nevertheless, he had built a machine that could play tick-tack-toe flawlessly. At Flinders University in Adelaide, South Australia, he created a computer language specifically for artificial intelligence and wrote programs that proved mathematical theorems. Joining the Stanford Artificial Intelligence Laboratory in 1977, he worked with Hans Moravec and other pioneers of that nascent field. He later moved across the country to become a professor at the Massachusetts Institute of Technology (MIT), where he also became director of the 230-person Artificial Intelligence Laboratory.
Brooks has published papers and books about model-based computer vision, path planning, uncertainty analysis, robot assembly, active vision, autonomous robots, microrobots, microactuators, planetary exploration, representation, artificial life, humanoid robots, and compiler design. His books include Model-Based Computer Vision (1984), Programming in Common Lisp (1985), and Cambrian Intelligence (1999). A 1997 documentary featuring him was called Fast, Cheap, and Out of Control, after one of his papers in the Journal of the British Interplanetary Society. The title reflects his controversial idea that robots could be made quickly and inexpensively and allowed to run autonomously, thus “out of control.”
Brooks starts Flesh and Machines by discussing the things that separate people from animals: syntax and technology. He describes a famously clever African grey parrot that appeared to put words together in new ways; closer analysis, however, showed that the bird did not exhibit syntax. Chimpanzees can appear to have rudimentary technology, but they have been observed for decades without showing...
(The entire section is 1899 words.)