The Ontario Council of University Libraries (OCUL) first launched its AI and Machine Learning Initiative in 2024. The initial phase centred on four exploratory projects and a capacity building program for librarians and library workers.
Project Progress
The four pilot projects, including the completed Government Documents and Audio to Text projects, have enabled OCUL and its digital service arm, Scholars Portal, to deepen understanding of:
- How AI can support the enhancement of accessible and discoverable materials.
- How to develop evaluation frameworks, tools, and processes to assess the accuracy, quality, and practical value of AI-generated outputs within academic library workflows.
- How to compare models, prompting approaches, and system configurations to identify fit-for-purpose solutions and opportunities for optimization and fine-tuning.
Lessons learned across the projects underscore the continued importance of human oversight and quality assurance, whether to address ambiguities in source materials or to verify the accuracy and reliability of AI-generated outputs. The pilots also highlighted the critical role of adequate computing infrastructure, iterative testing and parameter refinement, and comprehensive documentation in achieving scalable, sustainable, and reproducible outcomes.
ACE Remediation Project
The project team built an application to test the remediation features of AI on a corpus of books from the Accessible Content ePortal (ACE). The first round of testing focused on the quality of optical character recognition (OCR) and AI-generated alt text for images, and found the OCR quality generally good but the alt text generally unacceptable. The team is working on improvements to structural elements (e.g. chapters, headings, and footnotes) as well as an improved workflow for alt text. A new OCR engine (MinerU v4) was implemented that handles document structure much more reliably. Alt text extraction for images was moved to a separate pipeline engineered with the DIAGRAM accessibility principles in mind. These elements along with tables of contents will be the focus of a second round of testing in late 2026.
Enhancing Virtual Reference Project
Supported by two member-led working groups and central OCUL and Scholars Portal staff, the Enhancing Virtual Reference project continues to test the LibraryH3lp chatbot for use in the Ask the Library service. The first phase of testing was completed by members of the Chatbot Beta Tester Working Group with a Phase 1 Testing Report (PDF autodownload) published in July. A drop-in for Ask the Library chat operators was held in late July to share findings from the first phase of testing and working group members’ experiences with the knowledge base creation of the LibraryH3lp product. The Virtual Reference and AI Working Group recently finished their work on the Enhancing Virtual Reference Project Defining Core Values (PDF autodownload), which provides high-level ethical guidance to support the project and inform exploration of the benefits and drawbacks of the LibraryH3lp chatbot.
Audio to Text (Whisper) Project
Through our ongoing partnership with LibTech Insights, members of the Audio to Text (Whisper) Project Team presented a Choice webinar in August 2026. With a recording available, the webinar explored the potential benefits of AI speech-to-text tools and goes into detail around technical considerations and evaluation approaches for these workflows in academic library contexts.
Government Documents Project
The Government Documents project concluded over summer 2026 with a full Project Closure Report (PDF autodownload) published in September. The report details how AI-enabled OCR and metadata extraction techniques can significantly enhance the discoverability and accessibility of digitized government documents.
Capacity Building
Three capacity-building streams help foster OCUL member engagement and learning across academic libraries. Since August 2025, the Capacity Building Program has provided 18 professional learning offerings to more than 6,500 participants.
The Beyond the Algorithm Reading Club continues bimonthly featuring a selected topic based on OCUL member feedback. The curated readings, recorded conversations with guest experts, and discussion questions can all be repurposed so libraries can host a local version of Beyond the Algorithm at their institutions. The September 2026 installment focused on AI and Trust.
The AI Exchange launched in March, showcasing real-world applications of member-developed training and tools. The October installment examined integrating AI into first-year graded assignments. OCUL member library workers wishing to present future AI Exchange sessions are invited to submit a brief proposal.
Lastly, the AI Tools for Academic Libraries blog with Choice’s LibTech Insights remains a popular resource to learn about practical uses for AI tools in library contexts. Due to strong readership, OCUL and the LibTech Insights editorial team have extended the blog series to May 2027.
Knowledge Sharing
Focus groups were completed in June 2026, with the final AI and Machine Learning Focus Group Report recently published. A drop-in will be held on Tuesday, October 27, 1-2 p.m. Eastern Time to share and discuss the findings with the wider OCUL member community. Staff from OCUL member libraries can email ocul@ocul.on.ca for details to join the drop-in.
OCUL and Scholars Portal staff continue to be invited to share expertise on the initiative’s projects and the use of AI in academic libraries, elevating the profile of libraries as key partners in supporting responsible AI-enabled teaching and research. Recent examples include Improving Access to Government Documents with AI presented at the University of Toronto Libraries AI Community of Practice and Academic Library Workers Navigating Generative AI: Across Experimentation, Integration, and Resistance presented at the 2026 GAIL Conference.
For More Information
Visit the AI and Machine Learning section on SPOTDocs, the OCUL wiki, for detailed information about the pilot projects and capacity building programming.
Questions about the OCUL AI and Machine Learning Initiative can be emailed to Program Manager Kari D. Weaver: kari.weaver@ocul.on.ca.