Artificial Intelligence in the Archive: A Case Study on Transkribus in The Hayes Collection Enslavement Era Records

Accessibility of enslavement era records is an ongoing effort that has many challenges. With majority of records being entirely handwritten, the searchability of enslavement era records require time consuming and laborious work. Uncovering the lives and narratives of enslaved people, that were historically hidden in the archival record, is a necessary step in improving access for researchers, genealogists, and other collection users. This study aims to explore the use of Transkribus, a handwritten text recognition software, in identifying records of enslavement, using documents from the Hayes Collection at The University of North Carolina at Chapel Hill Willson Special Collections Library.