LTI Webinar Wrap-up: Should Libraries Let AI Models Train on Archival Collections?

Videos, resources, and more!

Posted on July 15, 2026 in Webinars

Authors:

LibTech Insights editor
A librarian training an AI on his archival data

It’s no secret that AI companies are hungry for data to train their AI models, and libraries have troves of unique data in their archives. Under what, if any, conditions should libraries agree to allow their holdings to train commercial AI systems?

Dr. Leo Lo, the Dean of Libraries at the University of Virginia and AI expert, takes on this prickly question. He introduces the UVA Archival AI Protocol, a decision framework downloaded over 6,000 times since its release, and the Statement of Shared Practice, a coalition commitment now signed by 17 institutions and growing. Together, they reframe AI training access as a disposition decision rather than an access decision, a distinction that changes how libraries evaluate risk, obligation, and community trust.

Expect to learn more about

  • Know how to distinguish disposition decisions from access decisions, and why that distinction determines everything that follows.
  • Have a step-by-step method for classifying AI-related requests, from vendor contract clauses to direct training partnerships.
  • Understand how the Statement of Shared Practice translates individual institutional decisions into collective action, and how your institution can join.

Stay tuned until the end to listen to the audience Q&A!

🎥 Watch now

đź•š Timestamps

  • 00:00: Webinar housekeeping
  • 01:51: TIE overview
  • 02:26: LTI overview
  • 03:07: Speaker introduction, agenda, poll question
  • 07:10: People stopped searching. They started asking.
  • 21:11: When an AI company asks to train on your collections — what do you say?
  • 30:45: The UVA Archival AI Protocol
  • 38:36: No institution knows what AI companies are asking across the field. Together, we do.
  • 47:40: Audience Q&A

đź“„ Resources

🗣️ Speakers

Leo S. Lo is University Librarian and Dean of Libraries at the University of Virginia and Special Advisor to the Provost on AI Literacy. He founded UVA’s Artificial Intelligence Literacy and Action Lab (AILA) and developed the Archival AI Protocol and Statement of Shared Practice, now signed by 17 institutions.

His CLEAR Framework for AI prompting has been featured in over 200 LibGuides across academic libraries worldwide, and his Universal AI Literacy Framework has been adopted and adapted at multiple universities and libraries. As ACRL President (2024–2025), he led the development of the ACRL AI Competencies for Library Workers.

His current research focuses on the AI answer economy – how AI systems are transforming the way people find, evaluate, and trust information. His writing has appeared in the Journal of Academic Librarianship, College & Research Libraries, The Guardian, and The Conversation. He holds an EdD from the University of Pennsylvania and a Postgraduate Diploma in AI from Oxford’s SaĂŻd Business School.

Subscribe to the LibTech Insights newsletter 🌟

Related Posts