John Sheridan (Digital Director at The National Archives) spoke at the second Unlocking our Digital Past workshop on 16 September 2021. John’s talk focused on three points of interest for archives and AI: i) what kinds of insights can be derived from AI approaches, ii) how should AI be deployed, and iii) how to select and preserve the AI tools we are using.
John’s abstract read as: “Digital archives hold important evidence. With the advance of machine learning techniques, the opportunities to gain new insights from data in the archive are rapidly expanding. Similarly, digital archives might also be used to synthesise new content. We do archives stand, as we find ourselves in the midst of the fourth — and perhaps beginning the fifth — industrial revolution, driven by AI? This presentation will address three challenges, from the perspective of The National Archives:
– What kinds of insights and value can be derived from digital archives, using AI based approaches? Where are the greatest opportunities? What potential harms should the archive protect against?
– How should AI approaches best be deployed to aid information management and the selection of records for the archive? What’s possible and where are the limits? In particular, how do we make sure that the value of records in aggregate is properly accounted for when making selection decisions?
– How do we select and preserve the AIs that are now such an important part of our digital fabric?”
You can view John’s presentation below: