Centered around a series of four exploratory projects and a Capacity Building Program that offers professional development opportunities to librarians and library workers, the Ontario Council of University Libraries (OCUL) moves forward its exploration of AI and machine learning’s responsible, ethical use in the academic library environment.
Project Milestones
Following the last update in August 2025, two of the exploratory projects, Government Documents and Audio to Text, have made significant progress.
The Government Documents Project continued testing various optical character recognition (OCR) tools, ultimately deciding to run the entire document set through Chandra OCR. The project also investigated using national compute clusters to speed up processing time, and successfully executed multi‑threaded OCR jobs across machines over the winter break. Work progressed on refining metadata outputs, aiming to include as much detail as possible while preserving a high level of accuracy. The project continued testing metadata across several models, including Ministral8b, GLM4.7 and Qwen 3-coder, adjusting prompts in response to field changes and model limitations. Development of the web‑based testing tool also moved forward, with new enhancements such as CSV export capabilities to support quality assurance across larger datasets. Project team members have been actively engaged in the Higher Education Quality Council of Ontario's Consortium on Generative Artificial Intelligence. In addition, the project team met four times with the OCUL Government Information Community to demonstrate progress, discuss government information-specific metadata nuances, and gather feedback on field definitions and desired outputs.
The Audio to Text project finalized a test corpus of materials in English and French, validated a custom rubric for evaluation of AI-generated transcripts, and recruited volunteer evaluators from across the OCUL Video and Accessibility Communities. The technical project team established a processing workflow and significantly improved the fine tuning of the Whisper automatic speech recognition model used for the project. Between October and December, the volunteer evaluators completed assessment of the entire test corpus and provided guidance on finishing refinements to the rubric. The project confirmed OCUL and Scholars Portal’s ability to set up a fully functional processing pipeline for AI audio-to-text transcription generation with a high level of accuracy. Volunteer evaluators believe such a workflow would save significant time toward the remediation of audiovisual materials. A project closure report with more detailed findings is available on SPOTDocs.
The other two exploratory projects in the OCUL AI and Machine Learning Initiative will progress throughout the remainder of this year. The Enhancing Virtual Reference Project, led by the efforts of two member working groups, is nearing completion of the first of a three-phase testing process. This initial phase investigates and documents the effort and ease of populating the LibraryH3lp chatbot knowledge base that serves as the basis for chatbot queries. The Accessible Content ePortal (ACE) project, which looks at improving the existing ACE workflow with AI tools to reduce the amount of remediation required at OCUL member libraries, will kick off in March 2026.
Building Capacity
The Capacity Building Program provided several professional development offerings throughout the fall and continues to grow in engagement and impact. In partnership with Choice’s LibTech Insights, OCUL launched the AI Tools for Academic Libraries blog series in August 2025. This bimonthly publication offers a clear, practical look at various categories of AI tools to help empower academic library professionals to make informed, confident decisions about AI technologies. Topics covered to date include AI research assistants, coding and programming tools, task management tools, and speech-to-text platforms building off the success of the Audio to Text project.
The Beyond the Algorithm Reading Club offers an opportunity to examine ethical dimensions of AI through a preselected reading, a recorded conversation with an expert guest, and virtual community discussions. The programming is offered bimonthly and has had strong interest and engagement through views and participation in community discussions.
New in 2026, the OCUL AI Exchange Program fosters cross-institutional learning and collaboration around AI in academic libraries. In this programming, individuals and teams from OCUL member libraries share locally developed AI-related training, programs, and expertise with colleagues across the province.
Lastly, this past December, OCUL and the Canadian Association of Research Libraries, together with University of Ottawa, Library and Archives Canada, and Carleton University collaborated to offer ARL/CNI’s Futurescape Libraries AI Toolkit Workshop in Ottawa. This training uses a scenario-based approach to help library leaders and staff explore future possibilities, test strategies, identify opportunities and vulnerabilities, and build readiness for long-term change within their organizations.
Impact and Awareness
OCUL and Scholars Portal staff affiliated with the AI and Machine Learning (AIML) Initiative have been working to disseminate the interest and impact of our work on artificial intelligence in academic libraries. In December, Jacqueline Whyte Appleby, Associate Director at Scholars Portal, and Stefania Kuczynski, AIML Metadata Specialist (co-op), presented Buried Treasure: Using AI to Increase Findability of Canadian Government Documents about the Government Documents project at Government Information Days 2025. Catherine Steeves, AIML Program Director, and Kari D. Weaver, AIML Program Manager, co-presented Collective Agency in the AI Era: Creating the Future Together at the 2026 Ontario Library Association Super Conference. Additionally, Kari was interviewed by the ARL Monitor about her work on AI Disclosure and co-authored a new book chapter on student motivation and generative artificial intelligence.
For More Information
For ongoing updates and detailed information about each exploratory project, please visit SPOTDocs, the OCUL wiki.
Questions about the OCUL AI and Machine Learning Initiative can be emailed to Kari D. Weaver: kari.weaver@ocul.on.ca.