From Archives to Insights: Integrating Transkribus and ChatGPT in Archival Workflows

27th June 15:20 – 15:40

Speaker: Dr David Brown

Abstract: In this presentation we will show what wonderful things are possible when you combine the output of one AI, Transkribus, with another, Open AI’s GPT4. Both AI’s harness the power of Large Language Models (LLM’s) that have become very popular and offer exciting possibilities for records management and outreach. Many heritage experts now use Transkribus, chatGPT or alternatives, but few as yet use them in combination. This presentation will demonstrate several use cases for workflows using AI, for example how to extract information from documents such as personal and place names, roles, and dates in various formats. We wil also see how GPT4 can proofread and correct Transkribus output and translate it to a variety of languages, creating a digital representation of the orginal text suitable for Named Entity Recognition tools. Finally, we will look at how to transform handwritten tables into readable slides, and covert 1,000 words into a picture!

Bio: Dr David Brown is a senior research fellow working on the Empire project at Trinity College Dublin and has worked on delivering a series of high profile digital humanities projects, including the Down Survey of Ireland (2013) and the Virtual Record Treasury of Ireland (2022). He is the author of Empire and Enterprise (2020), a study of early modern merchants in the Atlantic world, along with numerous book chapters and journal articles. A strong believer in the beneficial use of generative AI as a practical research tool, David has been demonstrating these new capabilities to librarians, archivists and historians over the past year.