LTI Webinar Wrap-up: AI Speech-to-Text in Practice

Videos, resources, and tips!

Posted on August 26, 2026 in Webinars

Authors:

LibTech Insights editor
Two librarians using a speech-to-text AI model

In her recent report on ALA Annual, Choice’s own Sabrina Cofer observed that AI conversation currently centers around what AI can do well and what it can’t. The ever-growing list of AI products (not to mention overhyped marketing) makes it difficult to adjudicate this line, which is why it’s important to have high-quality product evaluations based on real library use cases. Over the past year, our friends at the Ontario Council of University Libraries (OCUL) have helped us with this daunting task by sharing their research on current AI products and offering candid evaluations of their ability to meaningfully support library workflows. They have covered tools for research, coding, metadata, and much more, and we were thrilled to have the OCUL team give a webinar on their project for evaluating speech-to-text tools.

As the presenters demonstrate, speech-to-text tools offer a wide range of library benefits: from promoting accessibility to enhancing already automated (and highly tedious) workflows to increasing discovery. This webinar covers the breadth of their project and, more importantly, models for us the technical rigors and ethical considerations we should apply when evaluating any AI tool for library use.

🎥 Watch now

🕚 Timestamps

  • 00:00: Webinar housekeeping
  • 02:48: OCUL’s approach to AI
  • 06:39: Audio to Text (Whisper) Project
  • 14:49: The Evaluation Procedure
  • 24:01: Technical Considerations
  • 39:13: Testing Additional Models
  • 46:33: What We Learned
  • 51:08: Audience Q&A

📄 Resources

View presentation slides

Subscribe to the LibTech Insights newsletter 🌟

Related Posts