The future is about improving AI literacy, understanding these black box systems and making them transparent. – Interview with Sally Chambers

Sally Chambers is Head of Research Infrastructure Services at the British Library. She has previously worked at KBR, the Royal Library of Belgium and the Ghent Centre for Digital Humanities. She is also a member of the board of DARIAH (The Digital Research Infrastructure for the Arts and Humanities).

In this short interview, we talked about her role as Head of Infrastructure Services at the British Library, her PhD thesis, the role of AI in digital humanities research, archiving social media and her book recommendations.

Could you share some insight into your position at the British Library and what it entails?

The British Library is an independent research organisation, which means that it has a research remit, even though it is a cultural heritage institution. We try to ensure that the British Library’s collections can be used by researchers and provide access to research produced by British Library staff. Therefore, we have a research repository and that is where our staff can publish their research, publications, articles and conference proceedings.

We also have a portal EThOS – Electronic Theses Online Service, which was recently attacked, and we are currently working on restoring access. It is a portal where all the PhD theses awarded by UK higher education institutions are available. We also have a portal called DataCite UK where we support the creation and management of persistent identifiers such as DOIs (Digital Object Identifiers).

As part of DARIAH, we are working with our team on services and data policy and on a number of projects such as OSCARS and ECHOES. My favourite project is Collections as data, which is about providing access to digital collections of cultural heritage and making them available as data so that they can be analysed with AI, for example.

What is your role within the Centre of excellence for artificial intelligence in digital humanities (AI4DH)?

I am a member of the international advisory board. This is about sharing expertise and giving advice on how the CoE could connect to the various international initiatives and research infrastructures such as DARIAH or CLARIN.

Could you give us a glimpse into your PhD research in digital humanities. What are you investigating and why?

My research focuses on the question of what obstacles make it difficult for researchers to access cultural heritage collections. Cultural heritage institutions often invest a lot of time in digitising their collections, but then researchers use other collections and not the ones that are already available. My question is: how can we make the collections more accessible to researchers? Do researchers even know that these collections exist? Do they have the right format? Are they available as data? Often there are also legal obstacles – the data is in the digital libraries, but it’s not easy to get it out. In short, how can we improve collaboration with cultural heritage institutions such as national libraries and digital humanities researchers?

Do you have any case studies?

At the Royal Library of Belgium, we launched a project called DATA-KBR-BE, which focuses on a digitised newspaper collection. The goal was to explore how different types of researchers like historians, literary scholars, and media or journalism scholars use this kind of data.

For instance, historians were interested in strikes and instances of collective action that took place in Ghent during a specific year. They examined how these events were portrayed in the newspapers of the time.

Literary scholars, on the other hand, focused on something entirely different. They analysed how literature was published in serial form within newspapers—where each instalment ended on a cliffhanger, encouraging readers to purchase the next issue to find out what happened next.

Communication and journalism scholars were particularly interested in the history of Belgian journalism. They investigated which newspapers individual journalists contributed to. However, they often encountered a challenge: newspapers frequently listed only the journalists’ initials, not their full names. This led to a key research question: can we identify which journalist wrote which articles, and what topics they covered?

Do you use AI in your own research?

I’m a bit scared of AI. We did some work with machine learning and computer vision. We wanted to segment digitised historical newspapers. One of our PhD students at Ghent University used a combination of computer vision – where the articles appear visually – and other techniques to cut out the articles.

“Just as we archive newspapers to preserve a cultural and historical record, the same applies to social media. It is our mission to collect and preserve these digital materials for future research and cultural memory.”

You also mentioned the importance of social media archives for the digital humanities. Can you tell me more about this?

In Belgium, we worked on two related projects. The first was PROMISE (Preserving Online Multiple Information: Towards a Belgian Strategy), which focused on establishing a national web archive. The second, a complementary project called BESOCIAL, aimed to archive social media content.

As part of BESOCIAL, we crawled websites related to culture and invited the public to suggest social media feeds they believed should be preserved such as posts by politicians, for example.

Just as we archive newspapers to preserve a cultural and historical record, the same applies to social media. It is our mission to collect and preserve these digital materials for future research and cultural memory. However, providing access to social media archives poses significant challenges. Since the authors of these posts are often still alive, privacy concerns and GDPR regulations must be carefully considered before granting access.

How is your area of expertise evolving, and what should we be watching for?

This is a tough one. AI is very sexy, but we must make it usable for everyday work. I am a librarian and I am interested in library collections. Only a small proportion of historical newspapers are digitised and I am interested in how we can use AI to expand the collections. And of course, the future is about improving AI literacy and understanding these black box systems and making them transparent.

Do you have any book recommendations?

There is a new book called Feminist Digital Humanities, edited by Lisa Marie Rhody and Susan Schreibman. Jennifer Edmond has published a book on social science and humanities research infrastructures called Digital Technology and the Practices of Humanities Research. I would also recommend our forthcoming book Digital Humanities for Librarians, which needs some work but will introduce librarians to the digital humanities.