Designing Effective Speech Interfaces

Interview with Kate Dobroth

Usability Engineering Group

American Institutes for Research

B.A. in Psychology and Linguistics

M.A. and Ph.D. in Experimental Psychology and Psycholinguistics

Kate Dobroth is with AIR, the American Institutes for Research. Kate works out of the New England office, with the Usability Engineering Group. She has a B.A. from Wesleyan, in Psychology and Linguistics, and an M.A. and Ph.D. from Northeastern in Experimental Psychology and Psycholinguistics. Her studies have concentrated on how people comprehend spoken sentences.

Tell me about your current work.

Right now I'm working on the electronic assistant applications that all the big telephone companies are putting out. Something that will let you do voice dialing and will let you access information.

Tell me about what you think are some of the challenges for GUI interface designers. What are some of the things in the process or challenges in the interface that are different? What might throw them?

People have really high expectations of how speech recognition ought to work.

The thing that throws everybody, regardless of whether you are working on a handheld, a speech-only, or integrating a speech into a GUI, is that people have really high expectations of how speech recognition ought to work.

Too high?

The design challenge is really to figure out some way to constrain what people say.

Yes, definitely. ...