The presentation explores collaboration between the Louie B. Nunn Center for Oral History at the University of Kentucky and the National Library of Australia to research and experiment with speech recognition and text analysis technologies as applied to oral history transcription and access workflows in the archive. The National Library of Australia’s collection has 24,000 oral history interviews totalling over 50,000 hours, and the Nunn Center’s collection has over 12,000 interviews that total over 20,000 hours. Both the Nunn Center and the NLA created successful time-synchronization and search systems to enhance discovery and access to online oral histories that involve timed summaries and indexing as well as synchronized transcription. The search and user experience in both OHMS and AMADS is optimized when users are presented with both a timed-summary / OHMS index as well as a verbatim transcript, yet affordability of verbatim transcripts is becoming increasingly unattainable. Doug Boyd and Kevin Bradley will present the results of their experiences incorporating automatically generated text and text analysis into their accessioning and access workflows and explore the advantages and limitations of these technologies in the context of automation and AI, especially concerning archived oral histories.
