DAM’ed if You Do, DAM’ed if You Don’t: Using Generative AI to Support a Digital Asset Manager Migration R.C. Miessler Gettysburg College, United States of America
Gettysburg College's Musselman Library undertook a migration from CONTENTdm to AM Quartex, but the library did not have coding or development support to leverage the CONTENTdm API. Instead, ChatGPT was used to develop Python scripts to automate certain aspects of the migration, generally successfully.
Creating Text and Data Mining Collections for Small Institutions Mackenzie Brooks, Paula Kiser Washington and Lee University, United States of America
Though often associated with large digital libraries, rich text and data mining datasets can be created from local archival material at any institution with benefits for researchers and classrooms. This workshop will lead participants through the process of creating datasets as part of their regular digitization procedures with an MPLP approach.
“We Thought It Was One Tape, One Record”: Metadata Mayhem and Workflow Lessons from a Large-Scale A/V Digitization Project Louise Smith, Katie Ehrbar University of Southern California, United States of America
What do you do when your A/V collection has no metadata, AI transcriptions hallucinate wildly, and a single VHS tape contains nine discrete events? This presentation shares workflow strategies, lessons learned, and future recommendations from a year-long digitization and metadata effort on Jerry Brown’s personal A/V collection.
Allmaps: Open Source Georeferencing for Historical Maps Ann Hanlon(1), Ian Spangler(2) 1: University of Wisconsin-Milwaukee, United States of America; 2: Leventhal Map & Education Center at the Boston Public Library
Allmaps is an open source IIIF application that makes it easy to georeference digitized historical maps. This presentation will demonstrate how Allmaps works and how it can be used to extend the use of digitized map collections for research, and the implications for digital collections, GIS, and additional development underway.
From Archives to Access: AI-Powered Transcription for Multimedia Collections Matthew Butler University of Iowa, United States of America
This presentation introduces a validated AI pipeline for generating transcripts and summaries of historic political ads. Developed collaboratively and described in a recent peer-reviewed study co-authored by the presenter, the project demonstrates tools for evaluating and enabling access to video collections, offering librarians, archivists and developers practical methods to enhance discoverability and reuse.
Access in an Age of Censorship Amber Dierking Grand Valley State University, United States of America
The Queer Liberation Library (QLL) is doing information access differently. A fully digital, queer, nonprofit, library, QLL is accessible to anyone in the USA – for free. Using existing digital library tools and minimal infrastructure, QLL’s creative approach has had a transformational impact on LGBTQ+ library patrons’ access to information nationwide.
Metadata Augmentation Using AI: an IMLS Grant Project Jennifer Young, Jamie Carlstone Northwestern University Libraries, United States of America
Northwestern University Libraries is developing a generative AI/ML tool to create and/or augment item-level descriptions for digitized collection materials at scale thus allowing metadata librarians to provide deeper analysis and context. We will discuss testing and assessment of the tool as well as the ethical implications of the work.
Measuring and Critiquing genAI tools for Geographic Metadata Creation Erik Radio, Phil White, Nickoal Eichmann-Kalwara University of Colorado Boulder, United States of America
At the University of Colorado Boulder, we tested generative artificial intelligence tools’ abilities to accurately identify geographic features depicted in images from the library’s digital collections. This presentation reviews the results and presents them in the context of a critical analysis of similar tools.
Modernizing Access: Project Management Insights from the UCSF Industry Documents Library Website Redesign Melissa Ignacio University of California San Francisco
This session explores project management strategies behind the UCSF Industry Documents Library website redesign. It highlights how a remote, cross-functional team collaborated to select methods and tools that ensured a successful and user-centered redesign that meets evolving research needs, implements accessibility, and incorporates updated UCSF branding standards.
Borderlands Storytelling: Archival Collections, Digital Approaches, and Library Engagement in Innovative Humanities Scholarship Alana Varner(1), Verónica Reyes-Escudero(1), Megan Senseney(1), Chelsea Wells(2) 1: University of Arizona Libraries, United States of America; 2: Amherst College
This presentation explores strategies for facilitating community-engaged digital scholarship. We will discuss findings from a funded, three-year project at a research intensive, Hispanic Serving, Land Grant Institution that utilized library-based seed grants to engage researchers in digital and data storytelling efforts on and about the U.S.-Mexico borderlands.
Circa: A Customizable, Web-based Request System for Special Collections Jonathon Page North Carolina State University Libraries, United States of America
Circa is a web-based system for managing requests for items in special collections, developed by NC State Libraries. It integrates with ArchivesSpace, supports various request types, and utilizes IIIF. A new customizable, open-source version is in development. This talk explores Circa’s capabilities and invites feedback on its community-driven evolution.
