AI and Archival Practice On-Line Tutorials

Recorded: July 2024 

This session includes a discussion with Mr Jason Franks, Principal Data Scientist with RecordPoint.  Jason gives a demonstration of some of the AI capabilities that have been incorporated into the RecordPoint product, and also gives a preview of the work that is currently on the drawing board.  He demonstrates a machine learning capability for appraisal; used to inspect the content of records to assign a disposal schedule or category.  He also demonstrates the use of a LLM chatbot to produce summaries or descriptions of records, showing the power of the technology but also some of its limitations. 

Recorded: July – September 2024 

This video comprises interviews with three experts on various aspects of AI and Digital Preservation.  Ms Yaso Arumugam, of BlueRydge and former CIO of the National Archives of Australia, discusses the policy issues around AI in the creation of Government records and the impact this should have on digital preservation policies for Archival institutions.  Dr Hamad Al Mutairi of the National Library and Archives of the UAE shares his insights on the strategic considerations when applying AI to preservation in the context of national policies and global issues such as sustainability.  Dr Babak Hamidzadeh, University of Maryland Libraries, discusses the research he has undertaken within the InterPares Trust AI project, including the concept of ‘paradata’ to capture the wider context of AI in the creation and preservation of records and archives. 

Recorded: July – October 2024 

This video comprises interviews with three experts, sharing their experiences introducing AI to enhance public access to archival collections.  Ms Jiyoung Park, from the National Archives of Korea, is a lead member of an R&D project focused on designing an intelligent archive service model.  This involves an intelligent search that retrieves records for a researcher but also compiles a historical context for the records.  It also includes advancements for audiovisual records, such as voice-to text and facial recognition.   Mr Jean-François Moufflet and Mr Thomas Van De Walle, both of the Archives Nationales de France deliver a presentation of the SIMARA project.  This is the use of AI for the production of finding aids for the collection of the Middle Ages and Ancien Regime.  They explain their use of handwriting to text, named entity recognition, and transcription into EAD, resulting in finding aids delivered through a virtual reading room.  

Recorded: June 2024 

This video includes an interview with Mr Adrian Cunningham, a member of ICA’s Expert Group on Archival Description.  He discusses the current framework of international standards and work underway with descriptive standards.  He then explores the contemporary challenges facing archivists undertaking description with large volume digital collections and discusses the opportunities and risks around the application of AI for Archival Description.  

Recorded: June 2024 

This video includes a discussion with Mr John Balean of TopFoto.  John demonstrates the use of DataSheep, a software product developed for TopFoto by BlueBox.  Using DataSheep, John demonstrates how TopFoto is producing rich descriptions for photographic records based on the captions written on the backs of the original photographs.  DataSheep has incorporated APIs to use ChatGPT as a part of larger batch processing capabilities, enabling much greater efficiency in producing descriptions for large scale collections of photographic archives. 

Recorded: May 2024 

This video includes a discussion with two experts sharing their insights on the potential for AI to impact upon ethics and human rights when used for archival collections.  Ms Viridiana García Marignón, a national officer of UNESCO Communication and Information in Mexico, and Mr  Ernesto Ibarra Sanchez, President of the Mexican Academy of Cybersecurity and Digital Law, discuss a range of issues including: impact upon indigenous culture and languages, biases both in terms of data collected for AI as well as gender bias in the development of AI tools, the need for a ‘human rights’ approach to every AI project. They also provide advice on resources for ethical frameworks and codes of practice when applying AI to archival processes. 

Recorded: June 2024 

This video includes a discussion with Mr Tim Sears, Senior Software Engineering Leader at Groq.  Groq is an AI infrastructure company based in Silicon Valley.  Tim demonstrates some free or low-cost platforms that enable users to integrate their own business applications with publicly available LLMs and text to voice chatbots. 

In this set of three interviews, David Fricker speaks with researchers that have investigated the ways in which practising archivists and records managers should be thinking about AI for their professional practice.