News From UNC University Libraries: “With New Grant, “On The Books” Uses AI To Make Historical Records More Accessible”

From UNC University Libraries:

The University Libraries’ On the Books initiative is expanding its scope and exploring how artificial intelligence can make it easier to find and use materials from the archives.

A $765,000 grant from the Mellon Foundation will support three case studies using AI to improve access to materials related to groups historically underrepresented in institutional collections. Previous grants from the Foundation allowed the University Libraries to investigate text mining and machine learning as a way to identify discriminatory language in historical statutes.

“The Mellon Foundation’s generous support continues to help us apply new technologies to archival documents, and to advance our understanding of them as a result,” said Vice Provost for University Libraries and University Librarian María R. Estorino. “Using AI ethically and responsibly to make collections machine readable opens up research possibilities that go far beyond what an individual scholar or archivist could ever accomplish alone.”

Archives like Wilson Special Collections Library are a rich resource for learning about our past. But many archival materials are hard for users to locate and contextualize because they lack the transcriptions, descriptions or metadata that would make them easily searchable.

Each case study in On the Books: AI-Assisted Collections brings together library experts and users to address this challenge. The Library team is working with two historians, as well as community stakeholders and other scholars, to find responsible and ethical approaches to using AI in the archives.

• Historian Antwain Hunter researches firearm use by Black Americans in the antebellum South. He will work with a team to find relevant materials and transcribe them using AI, making them easier to use and access.
• Historian Monica Martinez is an expert in civil and human rights. She will help with the development of textual datasets created from Texas statutes, which will then be used to identify Jim Crow and Juan Crow laws.
• Community partners and scholars will help the team develop processes for using AI to create descriptions and metadata for historical photos of Black Americans’ everyday lives. That data will make it easier to find and understand those photos — especially for users with visual impairments.

“All of these projects build on the same idea that has driven On the Books from the beginning,” said Head of Digital Research Services Amanda Henley, who is leading the project. “We want to identify thoughtful ways of using technology to expand access to information about communities that have historically been overlooked in archival records.”

“This kind of work is only meaningful when it can be put to use,” says co-PI Matthew Jansen. “That’s why working with scholars is so important. We hope the real-world lessons from this project will eventually make it easier for other archives and researchers to use similar generative AI techniques with their own collections.”

Source

Filed under: Academic Libraries, Archives and Special Collections, Data Files, Funding, Libraries, News, Patrons and Users