Insights of the Search for Artificial Intelligence
By Dar-Lon Chang
Artificial intelligence (AI) has captured the human imagination since the beginning of civilization. Even our earliest ancestors dreamed of giving inanimate objects souls, and almost every religion speaks of the first people being created from materials such as blood and earth. Inventors and con artists throughout history have attempted to make machines that acted like human beings. It has only been in modern times that the dream of creating AI has seemed to be within our reach. The incredible developments of electronics and computers enticed scientists and engineers to explore the possibility of creating intelligent machines in our image. AI research began with a fury, but as the difficulties of understanding human intelligence mounted, the goals of AI research become less ambitious. Nevertheless, the precursors to true AI are part of our lives today, and as AI becomes increasingly more sophisticated, it will become even more imperative for us to face the social and philosophical issues raised by our attempts to create a new form of life.
Inspired by science fiction works such as Isaac Asimov's Robot novels, the first AI researchers began to characterize the nature of intelligence in the 1950s. In 1941, Isaac Asimov began writing stories of robots who had intentions, emotions and psychologies similar to those of humans. Naturally, scientists and engineers were impressed by Asimov's vision, and one of the first AI pioneers, Marvin Minsky, wrote, "Artificial intelligence is the science of making machines do things that would require intelligence if done by men." AI researchers used human intelligence as their model, but since human intelligence itself is difficult to define, their task of defining goals for AI was not trivial.
There were many objections to the possibility of artificial intelligence from the beginning. Some argued that only God can give souls to things, and others argued that thinking machines could spell our doom. This argument manifested itself in movies such as "2001: A Space Odyssey" and "The Terminator"; HAL and the Terminator ruthlessly destroyed their masters. Other objections included the idea that it was impossible to give machines consciousness and the possibility that humans have powers that are unavailable to machines, such as the Force in "Star Wars". Nevertheless, AI researchers continued probing the nature of human intelligence, and not unexpectedly, their work was influenced by the developments of psychology and biology.
Human intelligence is incredible because it can deal with the world in a "fuzzy" sort of way. People use "rules of thumb" to deal with daily situations even though rules of thumb are not necessarily logical nor guaranteed to work. These rules of thumbs are called heuristics, and the ability to use heuristics has been one of the most fundamental characteristics that have distinguished humans from machines traditionally. Machines can only deal with a controlled situation in a specified way. Humans, on the other hand, are capable of adapting to new situations using heuristics gathered from experience. Moreover, humans can learn from successes and failures and create new heuristics. AI researchers recognized the need for fuzzy logic mimicking human logic early, and they began to develop methods for incorporating the fuzzy factor. Fuzziness is important for a number of AI projects including pattern recognition systems, sensory research, language comprehension, game playing and problem solving. Without fuzziness, machines would not be able to take on human characteristics such as creativity and unpredictability.
AI has resulted in a number of limited successes. Several speech recognition systems with the ability to hold a limited conversation with a person have been developed. Vision systems such as POPEYE are capable of distinguishing the lines of an object and identifying them. Industrial robots capable of identifying objects in a bin and picking them out have been developed. A robot musician named WABOT-2 can sight-read music and play music with its articulated fingers on a keyboard and its feet on the pedals. Computer chess players such as Deep Blue have scored some wins against human grand masters, but World Chess Champion Gary Kasparov has still not lost a match to a computer opponent.
The most important practical development of AI has been the emergence of expert systems. Expert systems incorporate advanced information retrieval methods to mimic the abilities of a human expert. Expert systems have been used in medicine to help diagnose patients using statistical procedures, in civil engineering to help design bridges, in mathematics to prove theorems and in chemistry to identify molecular structures.
For all of AI's successes, though, true intelligence is still a long way off, and many people have doubts about the possibilities. AI research groups have been forced to scale back their ambitious plans. The incredible complexity of human intelligence has postponed the date of the arrival of an intelligent machine. Nevertheless, AI research continues, and machines continue to become more human-like.
Although the technical issues of AI are complex, the social and philosophical issues raised by the possibility of AI will be no less complex. As robots become more skilled and competent, people inevitably will be displaced. Even human jobs that seemed to be untouchable by machines in the past such as bridge design and medical diagnosis are endangered. Important political and social decisions will have to be made, and new jobs will have to be found for displaced workers. It is hoped that AI will create as many new jobs as it replaces, but nobody knows for sure. As robots and AI become more important in the workplace, safety and liability issues will emerge. To what extent should we entrust machines with our lives, and to whom do we assign responsibility when something goes wrong? Moreover, do we have full control over the machines? People are already fascinated by the possibilities of robots, and the emergence of robot characters in entertainment has reflected and reinforced that fascination. As robots become more like humans, will they turn against us, or should we give them "robot" rights? These questions and others will have to be answered as progress in AI continues.
Along with the search for grand unified theories in physics and the cataloguing of human genetics in biology, the search for artificial intelligence is one of the most important human endeavors of the present. As we learn more about AI, we learn more about ourselves. Not only are we learning more about the meaning and significance of intelligence, we are also shaping and forming our future relationships with our technological marvels.