Throughout my career I have always been interested in the intersection and overlap of ICT with archives. This has included digitisation, collections management including online catalogues, websites and content management systems and digital archives. I have always looked to be an archivist with an interest in the use of technology rather than looking to be both archivist and developer/coder. Whilst being broadly aware of AI, I had kept a mindful distance and not delved too closely.

Critical Cataloguing for Digital Preservation

This changed last year when I was commissioned by Dr James Baker at the University of Southampton to undertake the evaluation of its AHRC-funded project Critical Cataloguing for Digital Preservation. I attended the project’s event Preserving our Digital Pasts: Stories, Methods, Futures which successfully mixed reflection, discussion and training. The event also featured the first demonstration of the Digital Preservation AI Assistant an unexpected additional outcome from the project.

I asked it to list digital preservation systems being used in the UK, and whilst it provided a response I was acutely aware that the lack of any references prevented me from following-up or independently verifying the answer. If you know the answer the gaps are obvious, as some journalists discovered when asking DeepSeek about Tiananmen Square, but if you don’t they are not (which is worrying).

Without my own collections it hadn’t been possible to join some of the ongoing work across both the archive and museum sectors, but I am keen to seen and learn what might be/is currently being done. At the end of last year I attended The National Archive’s Annual Digital Lecture Turning over a new leaf: AI ethics in/through the archives [link is to a YouTube recording of the event].

AI and Digital Preservation: Good? Bad? Or just different?

So in the week that DeepSeek hit the headlines, the DPC event AI and Digital Preservation: Good? Bad? Or just different? was most timely indeed. It brought together a range of speakers with different perspectives and experiences to share their experiences:

• Adam Farquhar, Digital Lifecycle Management
• Paul Gooding, University of Glasgow
• William Kilbride, Digital Preservation Coalition
• Manuela Pallotto Strickland, King’s College London
• John Sheridan, The National Archives

Without summarising each presentation, though I’m sure there is a tool that does this, the key things that I took from the session were the following:

• there is both a high level of interest AND high levels of scepticism about the potential impact and contribution of AI
• as information professionals we had “an obligation to be involved” (Sheridan) and that we needed to understand what it is good at/for and only though engagement can we understand its limitations (in terms of what and why)
• ‘Responsible AI’ which considers accountability, transparency and compliance which are fundamental concepts in archives and records management
• ‘Slow AI’ which encourages learning about the process and reflection on the application of AI including the consequences of using it
• that some digital preservation suppliers are actively investigating the potential of AI
• as custodians of a large quantity of high quality data we should be looking to engage and collaborate with colleagues (Kilbride)

There are a growing number of projects that are already engaging with/using AI, a few that were mentioned in the session included:

• Our History Our Stories, University of Glasgow
• Using AI for Digital Records Selection in Government, The UK National Archives (Oct 2021)
• InterPARES Trust AI, multi-national initiative (2021-2026)
• a 3 part video series AI and Archives produced by Adam Farquhar (2024) in collaboration with the eArchiving Initiative

Conclusion

It was good to see a healthy discussion and one where ethical considerations feature at such a critical and fundamental level with confidence and trust being mentioned multiple times. I need to catch-up with some past projects, watch the ArkArchive videos. I will also look out for opportunities to stop watching and get more actively involved with something AI related this year and provide a more informed perspective.