Artificial Intelligence and Archives: What comes next?

Workshop 3 was organised by the University of Edinburgh (Centre for Data, Culture & Society) in partnership with the AURA project team.

The purpose of this third workshop was to bring together key actors in the archive ‘circuit’: from creators of data, to archivists and to users (thereby crossing the boundaries between Computer Scientists and Humanities Scholars) with the aim of planning new projects on AI and Archives. The workshop focused on the ethics of AI use in archives, looked at AI techniques such as machine learning, and reflected specifically on issues in digital humanities.

More than 150 participants attended the workshop.

SURVEY

We conducted a survey after the workshop and received 22 responses. The majority of these respondents were professionals from the GLAM [Galleries, Libraries, Archives, Museums] sector (45%) with an additional 15% identifying as GLAM sector civil servants. 18% of respondents were early-career academics (including MA/ PhD Students and ECRs within 6 years of their PhD), and 15% were mid-career or senior academics. The rest included ‘other civil servant/policy maker’ and university staff.

I) ORGANISATION. Respondents found the workshop very useful (64%) and useful (36%), with most saying the workshop was well-organised (91%) and the remainder (9%) saying the workshop was organised. Respondents found that the workshop was well-facilitated. In particular, the questions were well organised, the workshop seemed warm and engaging, fostered a “sense of community and positivity.” One respondent was grateful that “active participation [was] optional for those not confident in this area.” Three respondents praised the vision of the programme, with a fourth saying the timing of the workshop was good as starting later in the day enabled them to get some work done. Two respondents said the workshop kept to timings well, and another said that the breaks made the workshop a more enjoyable experience. Two respondents praised the accessibility of the online content, with one of these suggesting an online workshop could be preferable even once lockdown restrictions have been lifted.

II) CONTENT. Five people said the quality of the talks was the best thing about the workshop, two specifically praising the Keynote Professor Andrew Prescott for tying a range of different themes together. Fourteen respondents praised the range of speakers and various issues tackled in the workshop. In addition to this, four respondents said the workshop discussed topics that were relevant to a variety of fields. Comments included praises for the “range of speakers and professional areas,” the fact that the workshop broke down silos and was truly “cross-disciplinary and cross-sectoral,” and included speakers at “different stages of career” with “gender balance.” Respondents also appreciated the accessibility of the discussion, which was understandable by people from various backgrounds.

III) IMPACT. 18 out of 22 respondents replied to the question about the impact of the workshop on their own work. The majority of respondents said they had their view broadened or work developed in some way by the workshop. They said the workshop had made them more determined to go about their work, re-enforced their ideas, and gave them confidence to continue with the work in the way they were already doing it.

Comments included:

• “Very relevant to the work we do, particularly data sensitivity and born-digital management. Already working with academic institutions to look at ways of introducing AI technologies.”
• “I have been studying AI and audiovisual archives and have been too focused on the AV archives field. This broadened my view!”
• “As we develop our digital preservation capability this will help shape how we progress. It has reinforced my belief that we need to continue with discussions to become part of a project exploring AI.”

The workshop helped several respondents reflect on pedagogy in a fast-changing digital world. One respondent said: “The workshop really prompted me to think about what to teach (in anthropology and museum studies) especially in terms methodology – we are teaching old methods in a rapidly changing world. We need methodological innovations (and cross-degree elective course training components) to be up to speed with what collections and field research will (or really already should) look like.” Two respondents learned about a new project, and two took away some new reading from the day. Similarly, two respondents said they felt the workshop showed AI was inevitable and they needed to learn about it for their own line of work.

Involvement of non-AI specialists was presented as an action point to implement in the near future: “it is important that information rights professionals are involved in the development of AI within an organisation.” Another respondent mentioned that “the integration of AI into the archive profession is inevitable,” which shows the need for further engagement between archivists, AI specialists and Digital Humanists.

PROGRAMME

• 10:30 – 10:40 Welcome (Lise Jaillant and Melissa Terras)

• 10:40 – 11:40 Keynote presentation, “Can Archives make AI Better?”

– Professor Andrew Prescott, University of Glasgow

Chair: Bea Alex, University of Edinburgh

• 10 minute break

• 11:50 – 12:50 Focus on Digital Humanities

– Dr Larry Stapleton, “The INSYTE Cooley Laboratory, Waterford Institute of Technology: A Survey of Key Themes, Recent Research and Future Directions”

– Angeliki Tzouganatou, University of Hamburg, “AI, openness and participation in digital cultural archives”

– Dr Jennifer O’Mahoney, Waterford Institute of Technology, “The role of born digital data in confronting a difficult and contested past through digital storytelling”

Chair: Rachel Hosker, University of Edinburgh

• 1 hour lunch break

• 13:50 – 14:50 Focus on machine learning/ AI techniques

– Dr Giorgio Maria Di Nunzio, University of Padova, “Bias and Fairness in AI: New Challenges with Open Data?”

– Professor Matthieu d’Aquin, National University of Ireland, Galway, “AI for archives and collections: From processing metadata to analysing content”

– Bram van der Warf, “What comes next? AI for discovery or destruction?”

Chair: Rachel Foss, British Library

• 15 minute break

• 15:05 – 15:45 Focus on Ethics

– Dr Jenny Bunn and Mark Bell, The National Archives, “Archives and AI: What now?”

– Dr Adam Nix, De Montfort University, “Finding light in dark archives: Using AI to connect context and content in email”

Chair: David Canning, Cabinet Office

• 15:45 – 16:45 Roundtable discussion

Melissa Terras; Joe Nockels; Rachel Hosker; Mike Bennet; Kirsty Lingstadt; Anthea Seles

• 16:45 – 17:00 Short talk and closing remarks (Dr Annalina Caputo, Dublin City University)