Paradata at The National Archives of Norway

How should, and can, we document the use of AI? Part of the answer is paradata. Paradata, sometimes called process metadata, is used to describe the processes involved in creating data, including data created by AI. How is paradata important to describe the use of AI in the archival field, e.g. when machine reading by AI models, such as HTR, is used?

André Nilsson Dannevig and Inga Lang from The National Archives of Norway will present their work to create an ontology for paradata in the archival field, as well as practical examples of how paradata can be used to the benefit of archival users and the users of the AI themselves.

The goal of this presentation is to present the ongoing work and to gather insights from experts in the field of knowledge graphs.