The University is looking to apply subtitles and searchable closed captioning to all lecture recordings on the Canvas platform.
Close Captions and Subtitles are currently available via lecture recordings created via Zoom in the current climate of online learning. However, subtitles and closed captioning is not currently available for in-person lecture recordings on Canvas. According to the Universities accessibility website, students can currently access closed captions for recordings in English through Google Chrome’s live-caption function, released last year. This feature aims to auto-generate closed captions live as the video plays. Now, the University is aiming to work around this, by directly applying subtitles and closed captioning directly to Canvas.
The new measure is currently undergoing feedback through AUSA, who recently launched a survey asking students how the feature would impact students’ learning, and how much so. A significant piece of this form asks participants to rate, out of ten, how effective subtitles and closed captioning would be on an individual’s learning if they were 100%, 90%, 75% and 50% accurate. In order for subtitles and closed captioning to be completely accurate, lecturers or staff would need to proof and edit each sentence spoken in a lecture to ensure it is as accurate as possible.
Proofing is the main issue that was raised to AUSA by the survey respondents, as auto-generated subtitles and closed captioning are notorious for inaccuracies. The University of Minnesota states that Youtube’s auto-generated subtitles and closed captions are up to 60-70% accurate at best, with one in every three words potentially being wrong. In an educational context, particularly with non-english and/or complex words in the sciences, classical studies and politics spheres, incorrect words are also common. According to Rice University in Houston, auto-generated captions on Zoom are 80% accurate at its best, well below the level of accessibility needed. The institute states that transcript editing is needed to ensure learning equality for all students, particularly those who are hard-of-hearing.
Youtube had a community-captions feature implemented since its creation in 2005, but removed it in 2020, relying instead on auto-generated captions. This move was made to the dismay of many creators and members of the hard-of hearing community, who saw the ability as a way to better connect with the hard-of-hearing community, as well as translate captions into other languages. The backlash was so great that a petition was launched, which managed to gather nearly 50,000 signatures calling for the feature to return.
Ultimately, if the University wishes to add closed captioning, editing of transcripts will be needed. A reliance on auto-captioning results in inaccuracies and incorrect words, which would push back on the accessibility that the University is attempting to increase.