As institutions seek accessible, practical, and people-centered learning technologies, Automatic Speech Recognition (ASR) live captions are increasingly being explored as a scalable support tool in higher education classrooms. This session will cover a recent research study on ASR live captions in higher education and some practical demonstrations on implementing them in the classroom.
The study examined both the accuracy of ASR-generated live captions and student perceptions of their use in synchronous, in-person undergraduate learning environments. Two undergraduate general education courses were selected for analysis. Caption accuracy was evaluated using the Number, Error, and Recognition (NER) model, while students completed surveys measuring perceived usefulness, ease of use, and reliance, along with open-ended reflections on their classroom experiences.
Findings indicated that ASR live captions met the 98% accuracy benchmark suggested by the NER model, although serious captioning errors occurred more frequently than reported in previous studies. Overall, students reported neutral to slightly positive perceptions of the technology, with female participants demonstrating higher perceived usefulness ratings than male participants. Open-ended responses also highlighted both the practical benefits and limitations of live captioning in real-time classroom interactions. A practical demonstration of how educators can implement ASR live captions in face-to-face learning environments using readily available technologies and low-barrier strategies will also be shared.
Participants will explore classroom setup considerations, microphone and audio best practices, accessibility and ADA-related considerations, and ways to introduce captioning in a manner that supports inclusive learning environments. Attendees will leave with practical recommendations for integrating live captions into teaching practices to promote more inclusive, flexible, and learner-centered classroom experiences.
