What the Engineers would Like to Know from the Psychologists

Abstract

The emphasis in this NATO Advanced Study Institute has been largely on the techniques of automatic speech recognition and synthesis. There is increasing evidence that the years of research effort are now leading to some practical real-world applications and these are uncovering another set of problems which have received very little attention to date. At ICL’s Research and Advanced Development Centre, we are investigating how speech may be used for direct two-way communication between computer systems and their users. It has become clear to us that the technical problems of synthesis and recognition, important and difficult though they are, are not the critical problems that have to be solved in developing practical systems using speech. I would like to suggest that the critical problems lie in the human factors area. The main purpose of this paper is to indicate the nature of these problems and why they are so important, in the hope that this will cause more people to study them.