|
| |
Artificial Intelligence |
|
| |
|
|
| |
Artificial Intelligence (AI) is the most controversial area of computing today, an area beloved of sf authors—the duplication of human thought patterns by computer. It is an area in which much research has been done since the end of the World War II, beginning with the theoretical work of Alan Turing (1912 - 1954) in the 1940s.
Thanks to our limited understanding of human thought, it is quite difficult to define what the goal of artificial intelligence actually is. For example, it is not possible to answer the question ‘Do computers have knowledge of the data they process?’ unless you can define the term knowledge. (This question also has consequences in the legal field—can a computer be allowed as a witness in court if it does not really know what it is talking about?) This kind of question is not really answerable in the present state of philosophy and psychology.
It is also difficult to know how you will tell that a computer is managing to duplicate human thought patterns. A common idea is that this is achieved when a computer can do something that only human beings can. Many candidates for this ‘something’ have already been achieved: computers can learn from experience (how to play chess, for example), come up with completely new ideas (a computer known as the Artificial Mathematician came up with a result in number theory no one had ever dreamed of before), reason from insufficient knowledge (‘expert systems’ have been developed which will come up with answers in terms of probabilities to questions about mineral surveying and organic chemistry) and so on. With these systems which use incomplete or inexact knowledge, however, the possibility of mistakes made by the computer is introduced—something long thought to be a purely human domain.
The most famous statement of the goal of research in artificial intelligence is the Turing test, in which a computer and a human being are kept separate from the tester, and he or she asks questions the answers to which will tell the human and the computer apart. The test ends when the tester gives up or decides which is the computer. The computer passes the test if the tester gives up or is incorrect; it is then judged to have sufficiently human intelligence to be artificial intelligence.
No computer has ever yet passed the Turing test. Present goals in artificial intelligence include making computers understand English, respond to statements correctly and reject nonsensical ones, or using computers in fine art and music as part of the creative element rather than just as tools or instruments. Some seemingly easy tasks have proved remarkably difficult for computers; it will be some time in the future before the first robot can be built that can cope with even the simplest non-laboratory situations, such as a home or office environment. SMcL
See also computer art; creativity; electronic music; epistemology.Further reading T. Forester, The Information Technology Revolution. |
|
| |
|
|
| |
Bookmark this page:
|
|
| |
|
|
| |
<< former term |
|
next term >> |
|
|
|
|
|
| |
|
|
|
