AI and the visualisation needs of researchers using email archives

Abstract

Correspondence held in the collections of cultural institutions is an important resource for researchers, including historians, biographers, and social scientists. Email has been the dominant form of correspondence for the last 30 years and email archives are now being acquired by cultural institutions as a valuable resource for their collections. There are challenges in collecting, preserving and providing access to email archives, as with all born-digital materials, but also challenges in using email archives as a researcher. One tool that can help researchers use an email archive is visualisation. Visualisation has been applied to email for a long period of time but there has been little research into the visualisation needs of researchers using email archives. The preliminary outcomes of a research project investigating the needs of researchers suggests that visualisations will be useful to the researcher, who is always under the pressure of limited time, to work efficiently with the email archive. Artificial intelligence (AI) can play a role in preparing the email archive for visualisation. However, the research outcomes suggest that visualisations will be viewed cautiously by researchers who want time with the actual email archive to develop their research and they are unlikely to trust visualisations alone. The use of AI in the production of visualisations has the potential to add further barriers to researchers’ level of trust.