Students have no idea that this what they are getting into. They just want to know what is true. They don't want to hear one professor's viewpoint. But that is what they get every time. For this reason one university is quite different than another and every course in AI is different at every school. This is the fun part of teaching. Professors like talking about their own work and their own ideas. They love trashing their enemies. They love talking about the research they are doing. The question is: Is this what students came to learn? By and large I think it is not.

JB: How do university requirements get established?

SCHANK: Requirements get set in a university body, from general graduation requirements for a B.A. to PhD requirements in any field, by a committee. This committee represents various interests. When I became a member of the computer science department at Yale I noticed that in order to get a PhD in Computer Science one had to take a course in Numerical Analysis (NA) and a course in AI. I couldn't imagine a bigger waste of time for my graduate students than to take a course in numerical processing by computers when they were trying to build smart machines. One thing had nothing to do with the other. The requirement was there because of political compromise. No one knew what it meant to get a PhD in computer science so they simply required a little bit of everything of everybody. This could take an entire year out of a graduate student's life for no reason, but no one questioned it. I got rid of the requirement by making a deal with the top guy in NA. His students didn't need to take our courses and ours didn't need to take his.

Now, it turns out that deals like this are very hard to make in a university. The top NA guy was reasonable, and more important he was someone whose livelihood was not threatened by such a deal. When professors lose students they may find themselves in deep trouble. Nontenured appointments can be eliminated and tenured faculty may wind up teaching subjects they know little about. Unless your subject is very popular, the only way to keep teaching your favorite subject is to make it a requirement. Believe me NA was not popular, but it was well funded so the NA guys weren't worried. But there is no way to eliminate the Linguistics requirement at Northwestern short of a revolution. No one would ever take such courses otherwise. University requirements are about politics not education.

JB: Do the students understand this? What are their expectations?

SCHANK: Students tend to have the view that the university knows what is best for them and that if they follow the recommended course of study their lives will work out fine. I once had a freshman advisee who asked me what courses to take. This was at Yale where course requirements were quite minimal. I said the world was open to him and he should take what he had always wanted to know about. He told me this was no good. He needed to know what would get him ahead in life and that he thought I was pretty successful so he wanted to know what courses I had taken my freshman year. I assured him that I went to school in the dark ages when one had no choice at all and that he should be happy he lived in such enlightened times. He insisted. So I mentioned that I had taken Western Civilization, English Literature, Physics, Chemistry, and Calculus, all required of all freshmen at Carnegie Tech. Those are the courses he took.

Students expect that the curriculum set forth for them by the faculty is meant to help them get where they want to go after school. This simply isn't true. In computer science for example, the skills that will get students jobs include various programming skills that are used in industry. One might think that computer science departments around the country would make sure that all these employable skills are taught in their curriculum, indeed one would expect them to be the center of the curriculum. Sorry. Most computer science professors are not familiar with the commercial packages that are in use on a daily basis in industry and even if they happen to know them, they consider them to be of little intellectual interest. So, a computer science student will learn the mathematics involved in making calculations about what is computable, they will learn the theory of designing programming languages, but they will not learn much of what they will ever use in the real world. Computer scientists want their field to be a science and they want students to attempt to practice that science despite the fact that the students are there because they want jobs in industry.