Computational Bibliography research seminar

Tuesday 2 December 2025, 5-6.30pm

Samuel Alexander Building, A112

‘From provenance to identity: practical approaches to the use of AI in generating data about manuscript collections’

Huw Jones and Yasmin Faghihi (Cambridge University Library and Cambridge Digital Humanities)

In this talk we will look at two case studies where we used AI for the generation of information about manuscript collections. The first is the generation of structured data from a legacy catalogue of Persian manuscripts in the collection of Cambridge University Library, and the second is the production of a dataset from a recently published survey of Welsh manuscripts. We will assess the use of AI as part of a workflow which includes digitisation, traditional programming, and manual steps, covering areas such as prompt engineering, fine-tuning and the selection of specific AI tools for specific tasks. The overall purpose of the talk will be to look at how AI can be useful in accomplishing limited and well-defined tasks in larger workflows, and at AI’s particular strengths in accomplishing certain kinds of automated procedures.