Exploration of Automatic Speech Recognition for Deaf and Hard of Hearing Students in Higher Education Classes

Abstract

Automatic speech recognition (ASR) programs that generate real-time speech-to-text captions can be provided as supplemental access technologies for deaf and hard of hearing (DHH) students in higher education classes. As part of a pilot program, we implemented ASR as a supplemental access service in biology, statistics, and other courses at our university. To identify the benefits and limitations of ASR as an access technology, we surveyed 26 DHH students and interviewed 8 of these students about their experiences with ASR in their mainstream classes. Participants believed that ASR was beneficial despite the errors that ASR continued to generate; however, the accuracy and readability of ASR need to improve so that students can better access spoken information through ASR. This paper reviews points for researchers to consider when designing and providing ASR as a supplemental access service in educational settings.