Speech output systems and the Captain Kirk problem

Abstract

The Captain Kirk problem of a computer-based information system that can be freely accessed by voice and will provide unrestricted spoken output is someway off. This is particularly true of the speech input side. However, at the output level, work is well advanced on the production of high-quality synthetic speech. The intermediate level of input request understanding and generation of output information for speech is a fascinating area, combining methods from artificial intelligence, linguistics and computer science, and much can be expected from a combined application of these disciplines