This article concerns itself with what artificial intelligence (AI) may have in store for how science will be conducted and communicated in the not too distant future. It highlights some of the AI literature to show that computers have been successfully programmed to perform many of the procedures involved in scientific decision making and problem solving. It briefly reviews the ongoing debate over whether systems designed to exhibit intelligent behavior really think in the same sense that humans do. Finally, it reports on the possibility that more of the process of science will in the future be carried out by such systems as a result of research linking AI with progress in cognitive science.
Get full access to this article
View all access options for this article.
References
1.
1. Herbert A. Simon, The Science of the Artificial (Cambridge: MIT Press, 1969).
2.
2. Joseph Weizenbaum, “ELIZA—A Computer Program for the Study of Natural Language Communication between Man and Machine,”Communications of the Association for Computing Machinery, 9:36-45 (1966).
3.
The quotation cited here is taken from Grant Fjermedal, The Tomorrow Makers (New York: Macmillan, 1986), pp. 106-107.
4.
4. Avron Barr and Edward A. Feigenbaum, The Handbook of Artificial Intelligence (Stanford, CA: HeurisTech Press, 1982), 2:184-192, 267-278.
5.
For a synopsis of authoritative viewpoints on the subject of implanting commonsense knowledge, see the Panel Discussion in ibid., pp. 138-60.
6.
6. Robert N. Coulson, L. Joseph Folse, and Douglas K. Loh, “Artificial Intelligence and Natural Resource Management,”Science, 237:262-267 (July 1987).
7.
7. Barr and Feigenbaum, Handbook of Artificial Intelligence, 1:113-118.
8.
For instance, Research News, 37(5-6):5 (May-June 1986), reports that Manfred Kochen at the University of Michigan is developing PLAUSIBLE MATHEMACHINE, a program that will be able to learn geometric principles on its own. It will attempt to rediscover the Pythagorean theorem, for example.
9.
There are some AI programs in the nature of expert systems that may be said to perform original discovery rather than rediscovery. One called TETRAD supplies assistance in the area of statistical modeling. TETRAD automatically filters through large quantities of nonexperimental data in search of causal explanations, helping the researcher to select among the great number of alternative causal models that nonexperimental research can generate. A detailed description of TETRAD is contained in Clark Glymouret al., Discovering Causal Structure (Pittsburgh, PA: Carnegie-Mellon University, Laboratory for Computational Linguistics, 1986).
10.
10. Herbert A. Simon, “The Steam Engine and the Computer: What Makes Technology Revolutionary,”EDUCOM Bulletin, 22(1):2-5 (Spring 1987).
11.
11. Pagels, ed., Computer Culture, p. 139.
12.
12. Ibid., p. 151.
13.
13. Edmund C. Berkeley, Giant Brains or Machines That Think (New York: John Wiley, 1949).
14.
14. Pamela McCorduck, Machines Who Think (San Francisco: Freeman, 1979).
15.
15. Lew Kowarski, “Man-Computer Symbiosis: Fears and Hopes,” in Human Choice and Computers, ed. Enid Mumford and Harold Sackman (Amsterdam: North-Holland; New York: Elsevier, 1975), pp. 305-312.
16.
16. Peter J. Denning, “The Science of Computing: Will Machines Ever Think?”American Scientist, 74:342-346 (July-Aug. 1986).
17.
17. Raymond Kurzweil, “What Is Artificial Intelligence Anyway?”American Scientist, 73:258-264 (May-June 1985).
18.
18. Fjermedal, Tomorrow Makers.
19.
adapted from Margulis and Sagan, Microcosmos: Four Billion Years of Evolution from Our Microbial Ancestors (New York: Summit Books, 1987).
