Augmenting Archival Data: Using Ai To Enhance Discovery and Access Through Automated Data Extraction

As the volume of digitized archival material grows, the challenge of transforming scanned material into accessible, data-rich records has become increasingly valuable. This presentation explores the practical application of generative AI and named entity recognition (NER) models to enhance data discovery and access within digitized collections. Using a series of AI-generated Python scripts, this project automates the extraction of key metadata from OCR-processed archival materials including names, locations, organizations, and dates along with the generation of concise, human-readable summaries. By transforming digitized documents into structured, searchable datasets, these methods enable users to more effectively locate, interpret, and analyze historical information that was previously hidden within scanned pages. The session will demonstrate how AI-assisted extraction can improve metadata volume and quality while increasing accessibility and searchability. Special attention will be given to challenges such as model accuracy on historical texts, variations in OCR quality, and ethical considerations related to automation in archival description. This work positions AI not as a replacement for professional expertise, but as a powerful partner for data professionals expanding their capacity to champion data stewardship, accessibility, and discovery. Attendees will gain a clear understanding of how AI tools can be leveraged in library and archival environments to strengthen metadata creation workflows, support FAIR principles, and open new pathways for research and digital scholarship. By merging archival practice with machine learning, this presentation illustrates how data professionals can lead the next phase of digital transformation bridging past and present through intelligent data extraction and enhanced discovery.