The subtitle for the Dreyfus book captivated my attention: The Power of Human Intuition and Expertise in the Era of the Computer. Assuming intuition was based on experience and training, they did not touch on intuition in the large. By showing computers could not emulate intuition in the small, it was clear intuition in the large was computationally out of the question. Added emphasis appears in red, and my reactions are enclosed in a box: (Note 141)

Despite what you may have read in magazines and newspapers, . . . twenty five years of artificial intelligence research has lived up to very few of its promises and has failed to yield any evidence that it ever will. The time has come to ask what has gone wrong and what we can reasonably expect from computer intelligence. How closely can computers processing facts and making inferences approach human intelligence? . . .

Since this writing, Big Blue finally beat the world chess champion Gary Kasparov. However this was done with a totally rational, brute force use of computational power. The question remained - what relevance did the game of chess have to do the nuances of daily living and problem solving? My answer - absolutely none! Big Blue's achievement was insignificant in the larger human picture.

In short, we want to put the debate about the computer in perspective by clearing the air of false optimism and unrealistic expectations. The debate about what computers should do is properly about social values and institutional forms. But before we can profitably discuss what computers should do we have to be clear about what they can do. Our bottom line is that computers as reasoning machines can't match human intuition and expertise, so in determining what computers should do we have to contrast their capacities with the more generous gifts possessed by the human mind. . . .

As computational devices, computers were world champions when I first started working with them in 1962. My programming course at the University of Chicago with Robert Graves gave me a clear picture of what computers could do and do quite well. While I learned what a computer could do in ten weeks, it would take the rest of my life to acquire a hint of the intuitive potential of the mind.

We are not proposing to exalt the intuitive at the expense of the analytic abilities so highly developed in our Western culture. . . . The hoary old split between the mystical and the analytic will not do in the computer age, for neither pole of that often misleading dualism names the ordinary, nonmystical intuition that we believe is the core of human intelligence and skill. Further, we shall show that analysis and intuition work together in the human mind. Although intuition is the final fruit of skill acquisition, analytic thinking is necessary for beginners learning a new skill. It is also useful at the highest levels of expertise, where it can sharpen and clarify intuitive insights. . . .

Although I found a clear development of intuition in the small in the Dreyfus' writings, they avoided dealing with the more mystical intuition in the large. I would not find the soul of intuition here. But I did find a thoughtful philosopher and competent operations researcher who were no longer swept away by the unwarranted claims of the rational elite. They were voices in the wilderness calling attention to our misguided love affair with the computer.

Thus Stuart . . . now cautions people against making the same first step fallacy in operations research that I observed in artificial intelligence. He points out that while operations research had early successes in modeling operational problems in the military and industry . . . that is no reason to believe that the same mathematical modeling techniques can tell experienced . . . business executives whether to diversify their companies, or public policy makers how to allocate their budgets. Problems involving deep understanding built up on the basis of vast experience will not yield - as do simple, well defined problems that exist in isolation from much of human experience - to formal mathematical or computer analysis. . . .

My work conceiving and developing linear programming models and stochastic computer simulations demonstrated the essential prerequisite that any problem I attempted must be a clear cut decision situation devoid of even the most trivial non logical problem dimensions. For example, a simulation model to evaluate alternative computer configurations did not recognize that the department Director had known and worked with the sales representative from one of the computer vendors for ten years. That single factor would have more to do with the final decision than hundreds of hours of my computational wizardry.

A careful study of the skill acquisition process shows that a person usually passes through at least five stages of qualitatively different perceptions of his task and/or mode of decision making as his skill improves. Understanding the dynamic process of human skill acquisition provides the framework for our investigation of machine intelligence. Once we adequately appreciate the full development of human skilled behavior, we can ask how far along this path the digital computer can reasonably be expected to progress. . . . The five stages we shall lay out are called novice, advanced beginner, competent, proficient, and expert. . . .

The Dreyfus brothers set out to thoroughly understand how an expert acquired their expertise. Through a clear appreciation of that process, they were in a position to evaluate progress in representing expertise in a computer program. Of the five stages they delineated, a computer could emulate the first three and perhaps part of the fourth.

Society must clearly distinguish its members who "know how" from those that "know that." It must encourage its children to cultivate their intuitive capacities in order that they may achieve expertise, not encourage them to become human logic machines. And once expertise has been attained, it must be recognized and valued for what it is. To confuse the common sense, wisdom, and mature judgment of the expert with today's artificial intelligence, or to value them less highly, would be a genuine stupidity. . . .

When I was responsible for the upper division required computer course, I fell into the mystique of rationality. This came out through my amazing quotes about the computational power of computers. As my perspective matured, I complemented that with astounding data about the structure and complexity of the human nervous system. Until I changed my tone, I was part of the stupidity perpetuating the myth of rationality.

The chips are down, the choice is being made right now, and at all levels of society computer type rationality is winning out. Experts are an endangered species. If we fail to put logic machines in their proper place, as aids to human beings with expert intuition, then we shall end up servants supplying data to our competent machines. Should calculative rationality triumph, no one will notice that something is missing, but now, while we still know what expert judgment is, let us use that expert judgment to preserve it.

A recurring science fiction theme dealt with the outcome of the progressive encroachment of rational intelligence into society. In these stories, a time came when humans were slaves to their computer masters. This occurred so insidiously that no one was aware that it was happening. The hero or heroine was one who realized what had taken place and fought back to recover their human dignity.

BACK