Source: Feurzeig, W., & Papert, S. (1969). CAI Problems and Prospects Spring Joint Computer Conference, 1969.
Note: This is an abstract for a conference panel discussion led by Patrick Suppes.
CAI problems and prospects
by Wallace Feurzeig
Bolt, Beranek and Newman Inc.
Cambridge, Massachusetts
and
Seymour Papert
Massachusetts Institute of Technology
Cambridge, Massachusetts
The expression “computer-assisted instruction” (CAI) is generally used to describe situations in which the computer is used as a teaching surrogate in some sense—whether as drill instructor, tester, or specialized tutor. Most applications of this kind have had specific, limited, and modest educational goals. When used to administer drills or branching tests, a computer is not called upon to be intelligent—only useful. Yet it is interesting and important to ask whether the computer can become an intelligent artificial teacher, and more generally, whether there are valuable ways of using computers for teaching and learning.
In artificial teaching, the computer controls the interaction with the student. There are applications of the opposite kind, where the student controls the machine. The most common one is the teaching of computer programming itself. Another is the use of a computer to simulate a “real” laboratory. And, there is a potentially rich spectrum of intermediate arrangements—strong instructional interactions in which the student and the computer share control and direct each other. As an example, student programming and artificial teaching might be coupled by having the computer monitor a student’s work as he uses a programming language to perform a simulated experiment or to solve a problem. No significant experiments in this direction can be done without a great deal of work; but we do know, in principle, how to make a program follow the steps of a student who is not constrained by a stereotyped pattern and how to diagnose his difficulties on the way.
We shall argue that computers will make deep contributions to education in all three areas:
- 1. first, and with capabilities already well established, through the teaching of programming languages;
- 2. ultimately, and to an extent largely dependent on progress in artificial intelligence research, as an artificial teacher;
- 3. intermediately, as an instructional monitor or assistant, in a number of different subjects as diverse as music, language, and physics.
Along the way, we shall elaborate on specific educational contributions including the following.
- 1. The teaching of programming can provide a conceptual and operational framework for the teaching of mathematics.
- 2. Using an appropriate language, programming can be introduced routinely to third-graders for its special value in teaching the skills of clear and precise thinking and expression.
- 3. The computer can enhance the teaching of “practical” subjects (such as navigation or speaking a foreign language) whose mastery requires the integration of mechanical and intellectual skills.
Finally, we shall contrast the present lack of depth and perspective characterizing much of the work in this field with its rich prospects. In particular, we shall discuss our view that a serious investigation of the problems involved in developing an intelligent teaching system will yield rich results in the fields of computers, education, and psychology.