Accessibility workflows are a crucial step in the full lifecycle of digital collections work. Too often, however, these workflows are time consuming, tedious, and a significant burden on teams with limited resources. How can we lighten the workload of creating accessible collections, while ensuring the accuracy of the information provided? While it may not be the answer to every problem, artificial intelligence (combined with human quality control) can help teams improve the accessibility of their digital collections. This improved access to collections will enrich the teaching and learning of faculty, staff, and students across campus.
In this presentation, presenters will explore the intersection of AI, accessibility, and digitized special collections. By exploring image descriptions, transcriptions of hand-written materials, and Retrieval Augmented Generation (RAG) for information retrieval, the team will share the ways in which AI has helped (or harmed) digital collections workflows.
Learning Objectives:
Explain how AI tools can be integrated into accessibility workflows for digitized special collections, including its potential benefits and limitations when paired with human quality control.
Evaluate the effectiveness of AI-assisted tools for image description and handwritten text transcription in improving access to digital collections.
Identify practical strategies for using AI to reduce staff workload while maintaining accuracy and ethical standards in accessible digital collections work.
Presenters:
Sidney Gao, Interim Director of Digital Initiatives, Digital Collections Manager, University of Cincinnati Libraries
Sean Crowe, Digital Projects Librarian, University of Cincinnati Libraries
Location
Setting: Live Virtual Online via Zoom UNITED STATES