OCUL Artificial Intelligence & Machine Learning Program Kicks Off

By Jacqueline Whyte Appleby

The Ontario Council of University Libraries (OCUL) Task Force on AI/Machine Learning started with a clear-cut premise: all  21 university member libraries should harness their collaborative governance and infrastructure to experiment with AI. The potential for exciting and useful end products was there, but the task force process itself would be valuable in helping OCUL build the skills and competencies to support the next wave of teaching, research, and innovation, both collectively and at our institutions.

However, transforming this intriguing, high-level, and open premise into a structured, funded, and staffed program was no easy task.

Supported by Mike Ridley, Librarian Emeritus at the University of Guelph, the Task Force led broad consultations, speaking with staff at many member libraries, Scholars Portal, the Canadian Research Knowledge Network, the Canadian Associate of Research Libraries, and Library and Archives Canada. They also met with a series of private-sector AI experts and some American university colleagues working in this area.

McMaster hosted an OCUL-wide Hackfest to facilitate applied learning using AI tools. Over 90 people participated, working in small groups to explore proposed use-cases and identify curiosities, needs, and potential obstacles. The task force released an interim report in March detailing the findings of this event. After months of discussions, a series of distinctive projects began to take shape.

This May, the OCUL Directors approved the Artificial Intelligence/Machine Learning Report and Strategy, along with two years of funding to execute five projects:

• Building an audio-to-text transcription pipeline
• Exploring AI to support virtual reference
• Enhancing government documents metadata
• Improving accessibility with remediation tools
• Building capacity across OCUL and with other national and regional partners

Each project has a specific objective that we hope to accomplish in the next two years. There is also a broader goal focused on skill development and laying the groundwork for future projects.

One example includes exploring audio-to-text transcription AI tool Whisper. Many OCUL members are interested in using it for rapid closed-captioning, enhancing video and audio file searches, and helping researchers with interview data. Rather than schools, libraries, or individual research teams needing to set up and maintain a Whisper server, Scholars Portal will host the tool for all of OCUL, allowing for quick access on request. Figuring out a Whisper workflow is the initial goal, but we recognize there will be other AI tools that staff want to experiment with or use.

However, some questions have come up: How will we identify and prioritize appropriate consortial tools? How much time and money will it take to maintain those tools? What will the procedure be when a tool is depreciated or no longer of interest? Getting Whisper up and running is a way for us to start addressing these questions.

These projects will be supported by Scholars Portal and OCUL staff, as well as two new committees. The AI/ML Compute Committee will advise on infrastructure resourcing, architecture, and technology needs, while the AI/ML Training Committee will identify, secure, and support training opportunities for OCUL members and other participating partners.

Direction and leadership for the program will come from the Program Director Catherine Steeves (Western University), the Technical Manager, Pieter Botha, and the soon-to-be-hired Program Manager, as well as the Directors of OCUL and Scholars Portal, all of whom will report to the OCUL Executive. A Data Analyst at the University of Toronto will also be hired to support software development and integrations.

We’re just getting started and we’d love for you to follow along! You can read the full report online.

—

Jacqueline Whyte Appleby is Associate Director of Scholars Portal. You can contact Jacqueline at jacqueline@scholarsportal.info.