Personal Computers Have Arrived, but Compared to What's Coming, Says An Expert, They're Still Tortoises

updated 12/17/1984 AT 01:00 AM EST

originally published 12/17/1984 AT 01:00 AM EST

It's handy! It's dandy! It teaches! It budgets! It plans! No home should be without one! What is this miracle appliance, this plug-in panacea? Why, a personal computer, of course. Or so many people have come to believe. Roger Schank, however, is not one of them. Schank maintains that it is still too early for the consumer to jump on the computer bandwagon, an argument he develops in his new book, The Cognitive Computer (Addison-Wesley, $17.95). As a professor of computer science and psychology at Yale University and director of its artificial intelligence project, Schank, 38, is not an impartial observer. He is intimately involved with developing the generation of "smart" computers that he says will replace today's relatively inflexible models in the not-too-distant future. And his two New Haven software firms, Cognitive Systems, Inc. and Compu-Teach, Inc., develop programs utilizing artificial intelligence technology. Schank is recognized as one of the most astute and imaginative computer scientists in the country. In his Yale lab, he talked with Correspondent Lynn Schnurnberger about how long it might be before the average computer becomes a simple, useful tool for the average person.

Why are you telling people to think twice before buying a computer?

Computers are in this funny phase where you have to know how to operate them in order to get anything out of them. In about five years, you won't have to know a blessed thing about them. Computers should be like your television set; you don't have to pore over instruction manuals or know what goes on inside your TV to get it to work. If you can't use today's computers without pain, then just wait. It's the computers that will have to change, not you.

Aren't there some things computers do well right now?

Sure. Word processing is one. Anyone who does any writing at all is insane not to have a computer, because they make life so much easier than a typewriter.

What about using a computer for home budgeting or small-business tasks?

Home budgeting you could do just as easily on a piece of paper. A small business, though, should have a computer. Good software has been developed for that kind of use. But I don't think there's a serious use for computers in the home today. That's my main point. The applications available other than word processing are uninteresting—like using a computer to balance your checkbook. I just don't see the value in it.

Yet people are rushing to buy computers. Is there a backlash building?

Statistically, I don't know. But I meet people on airplanes and at cocktail parties all the time, and they tend to tell me about their computers. Most are either very frustrated or they get so into it that they practically become mini computer scientists. And many admit they're enjoying the machines but not saving any time over the old way of doing things. Then there's a group who say they're nervous that they don't have one and wish they did. What they really want is to be assured they don't have to buy one. And, of course, they don't.

But are we risking our children's futures if we don't buy them computers?

One of the computer companies runs a commercial showing a bright, eager youngster going off to college. In the next shot we see him looking miserable and disheveled, having been sent home from college for "lack of computer skills." Now this is ridiculous. There will always be some people who know how to program—that is, create software—just as there are some people who know how to fix a car. The people who liked doing math in high school are the ones who are going to find programming fun. But you shouldn't have to know how to program. As for the educational software currently available, most of it is just drill and practice, often with a bang-bang video-game format. It's not very creative.

You contend that things will change in about five years. What can we expect?

We can look forward to computers that will respond to questions typed in plain English with answers displayed in plain English.

Would you give an example?

Some of the technology is here now. For instance, data bases—computerized libraries, essentially—already exist all over the country for different types of information. To gain access to them on a computer today you have to know the right access codes and the right query language. Let's say you want to know what kind of wine to serve with the dinner you're making and what's available locally in a certain price range. You don't want to bother with access codes. In five years you'll be able to just type in your question and get a detailed answer. The computer will pick the right data base—and in five years there will be more and better data bases—and get the information by itself.

What role will artificial intelligence technology play in this ?

The Al part is getting the machine to understand what you type—to understand plain English, in other words. It's tricky. What we've done is study how people understand English. Then we take what we learn and try to break it down into a set of step-by-step instructions for understanding English that can be fed into a machine. We have some ability to do this now, but it doesn't mean you can ask the machine anything you want. There's a big difference between a computer understanding a question like "What wine should I have with dinner?" and a metaphysical question like "What should I do with my life?" Although in five years computers will be able to field many practical questions, you're not going to see the development of a machine that can speak English profoundly for quite a long time.

What does artificial intelligence promise in the long run?

Ultimately we want to be able to get machines to learn from experience and make judgments, not just store and retrieve data. For instance, in my laboratory at Yale we are building programs that model little areas of knowledge. I have one guy who works on cooking, another on chess, another on horse racing, another on cancer. As we put more information about these subjects into the machine, the program tries to use the information to come up with hypotheses about the subject. We're trying to figure out what methods humans have for learning, and then we'll put them into the program. It's an experimental thing. Who knows? We may not succeed. ?

From Our Partners