Abstract
This presentation examines optical character recognition (OCR) through the lens of archival ethics as outlined in the Society of American Archivists (SAA) Core Values Statement and Code of Ethics, given the current debates surrounding artificial intelligence (AI). A historical overview highlights persistent challenges of authenticity and integrity, transparency and accountability, access and equity, and responsible stewardship and sustainability, as well as new concerns about bias, sustainability, and accountability using large language models (LLM). A case study describes systematic testing of LLM, transformer model (TM), and neural network (NN) architectures and examines the challenges in creating a reliable, scalable in-house OCR tool named Opticolumn. This case study finds that NN approaches better align with archival ethics than do LLM tools, which may generate fabrications, but that OCR tool choice will depend on the capacities and preferences of individual institutions.
Biographies
Rebecca Hastings is an Assistant Professor in the University of Idaho Library and a digital archivist with Special Collections and Archives. Her current areas of interest include AI in archival work; documenting student life in archival collecting and digital collections; and diversity, equity, and inclusion in archives. Previously, she was involved in academic history, specializing in Russian and Soviet history.
Andrew Weymouth is an Assistant Professor with the University of Idaho and the digital initiatives librarian for the Digital Scholarship and Open Strategies department. He focuses on using static web hosting to curate the institution’s special collections and develops digital tools and workflows to enhance transcription, tagging, and image processing to make the university’s media, text, and visual resources more discoverable for researchers. He is completing an MA in history at U of I under the supervision of Dr. Rebecca Scofield, focusing on pageantry and speculation in the Inland Empire.
Content: CC BY-NC-ND 4.0 Rebecca Hastings & Andrew Weymouth 2024 (get source code).
Theme: Variation on workshop-template-b by evanwill
