Computer 'crystal ball' gazing: Washington's OTA weighs pros, cons

If you have glimpsed the future and it was on a computer printout, you might be interested in a recent report by the Office of Technology Assessment (OTA).

The congressional research arm, in its report ''Global Models, World Futures, and Public Policy,'' takes a hard look at the increasingly sophisticated computer models of the world's physical, economic, and political systems. These are being used more and more to predict the shape of things to come.

Groups such as the Club of Rome have used global models, like that in the precedent-setting ''Limits to Growth'' study in 1974, to dramatize their efforts and points of view. National governments and organizations like the UN are leaning on these techniques more and more heavily as part of their planning efforts. In addition, a number of large corporations have also taken to computerized crystal ball gazing.

''Global modeling studies have varied widely in their purposes, techniques, findings, and prescriptions,'' the OTA researchers report. ''The results of some studies have been guardedly optimistic, while others have been highly pessimistic. . . . Nevertheless, they have generally identified the same problems and seem to have arrived at roughly similar qualitative conclusions about the present state of the world and its plausible futures.''

These areas of agreement include:

* Population, property, and wealth cannot grow indefinitely without eventually causing widespread hunger and resource scarcities. However, with improved distribution, there is no reason why basic human needs cannot be met for all the people on Earth.

* The continuation of ''business as usual'' around the world will result in growing environmental, economic, and political difficulties. Technological progress is essential, but basic political changes are necessary to avoid severe problems.

* Earth's future will be fundamentally different from its past. The next 20 to 30 years represent a major transition period. The outlines of this future condition are not clear or predetermined but are being shaped by decisions that are being made today.

* The people of the world are far more interdependent than widely realized. As a result, short-term, competitive, and strictly nationalistic policies are not likely to be as successful as expected. Cooperative and long-term approaches to world problems are more likely to benefit all parties involved.

* While global economic collapse due to pollution or resource depletion may not be a realistic concern, severe regional problems of global importance, such as starvation in South Asia, are very likely.

Some computer models of this sort have been used to dramatize much more extreme predictions of the future, but these were the result of often questionable assumptions, the report finds. And, while the modelers agree upon this general outline of the world's condition, they are sharply divided on solutions. These tend to fall along ideological lines: Club of Rome and environmentalists advocating limiting of population growth and economic development; UN and third world modelers emphasizing redistribution of wealth, for instance.

Just because a mathematical model of certain aspects of the world economy and society is developed, put into a computer, and run into the ''future'' does not mean that it can predict the course of actual events, the OTA researchers point out. However, such models can be valuable in estimating the odds that certain things are likely to occur given certain specific assumptions about world trends , national policies, and future events, they conclude. Like any other analytical tool, global models have certain strengths and weaknesses.

According to the report, these advantages include: a long time horizon (20 years or so), an ability to handle a large amount of information, the clear statement of assumptions and procedures which can be debated, precise mathematical logic, and the flexibility to explore a wide range of assumptions and policy alternatives.

On the other hand, these models involve a considerable simplification of the complexity and simplicity of the real world, and there are no generally accepted tests of a model's realism. There are a number of cause-and-effect relationships , particularly in the environmental and sociopolitical realm, which are little better than guesses. Also, in many areas, the statistics upon which such models rely is inadequate, unreliable, or inconsistent.

Currently, various agencies in the US government use computer models in planning, but all too often they use inconsistent and contradictory assumptions, OTA reports. This was highlighted by the Global 2000 report which tried to integrate these various models with limited success. If the United States is to reap the benefits of this evolving planning tool, such an integration is necessary, the report concludes.