2024 Oral History Association Annual Meeting

Abstract
Dartmouth College, in collaboration with the University of Kentucky, is embarking on a two-year project (2023-2025) to upgrade the Oral History Metadata Synchronizer with data annotation and data visualization capabilities drawn from the field of digital humanities. During this session, the project team will share lessons learned from our ongoing work, which has brought together oral historians, digital humanists, archivists, librarians, and undergraduate students to envision opportunities for enhancing access to oral history collections. We will discuss the history of this cross-disciplinary endeavor; engage in dialogue regarding key ethical and methodological questions that guide our current work; and invite audience members to imagine with us future possibilities for applying Natural Language Processing (NLP) tasks to oral history collections. Using OHMS as a case study, this session will engage with timely debates regarding the role of AI tools in the field of oral history. Audience members will have an opportunity to view a prototype of the OHMS application and a demonstration of its NLP functions. These examples will serve as jumping-off points for discussion about the benefits and limitations of automated data extraction methods when applied to oral history collections. The presenters will address and invite audience engagement around questions such as: How can we use the data extracted by NLP to enhance the accessibility of collections and provide new avenues for research? What safeguards and practices should we put into place to address inherent biases, inaccuracies, and privacy concerns? How can we balance automated efficiency with the importance of human review, deliberation, and care in the description of oral history collections? Through such discussions, the presenters aim to contribute to an ongoing, collaborative community of practice around the future of oral history collection and use in the age of AI.

Enhancing the Oral History Metadata Synchronizer - Douglas Boyd University of Kentucky
This presentation will share recent updates to the Oral History Metadata Synchronizer (OHMS), including its migration to the Aviary Platform and the creation of a new transcript editor. These developments will pave the way for a future version of the OHMS application with powerful capabilities in the areas of data annotation and visualization.

A Digital Humanities Approach to Oral History Collections - Edward Miller, Dartmouth College
This presentation will provide a history and overview of an NEH-funded collaboration between OHMS and the Dartmouth Digital History Initiative (DDHI), bringing digital humanities methods into the practice of oral history collection management. The project will result in a new version of OHMS that offers accessible and interactive avenues for scholarly research and teaching using oral history archives.

Natural Language Processing for Oral History Collections - Sanjana Raj, Dartmouth College
This presentation will introduce two Natural Language Processing (NLP) tasks - Named Entity Recognition and Named Entity Linking - and share practical examples of how they can be applied to oral history collections. We will also discuss advantages and challenges of using NLP tools to extract and annotate data in oral testimony.