INTERVIEWER: Gökhan Çolak INTERVIEWEE: Melissa Terras

Cultural Memory and Technological Mediation

How does your work in digital humanities redefine the preservation of cultural memory in a time of technological acceleration?

Its important to distinguish between preservation, and data analytics: digitisation is not preservation and digital media are very fragile. But in making digital copies or representations of cultural heritage, you can often do analysis on the digitised version which was not possible with the physical. You do need to look after the digital data carefully, and colleagues in digital preservation are explaining how best to look after digital assets long term.

Its easy to think, given digital is all around, that things will last forever, but what we are seeing here is that people know less and less about how fragile the digital environment is. So the two go hand in hand – make digital versions or copies, but look after them properly, to increase access to cultural heritage, and allow analysis that could not be done otherwise.

Context, Ethics, and Representation

What challenges arise in ensuring that digitization efforts remain contextually faithful to the historical and cultural specificity of materials?

We’re very lucky in that we are now past the experimental phase of digitisation. Back in the 1980s and 1990s, a dedicated bunch of librarians and archivist and photography technicians figure out the limitations and best practice of digital capture, as well as the opportunities. The most common guide now followed are the FADGI guidelines https://www.digitizationguidelines.gov/, which are often updated, and provide clear guidelines on how to digitise things properly – ie accurately while documenting the object AND the process of digitisation. This is critical – you should be able to know how something was made digital, and how use of the digital object is licensed, to be able to trust it – and then to use it in a way which is faithful to the original. Colour reproduction, for example, is something we know how to do well now, but you have to put the effort in, and use processes to ensure that the blue of a painting you see in real life will look like the blue of the painting once it is photographed and put on a monitor for others to see.

Considering current debates on algorithmic preservation, what role should AI play in managing and interpreting cultural heritage?

There’s lots of good that AI can do (and is doing) in cultural heritage management. Metadata creation, metadata standardisation, and handwritten text recognition are some of the mature AI uses in cultural heritage, which can speed up processes, and can be used very ethically, where it is clear that the data is that has been transformed by AI. All AI is not generative AI, although recently people think that. I think its very important to clarify that. The jury is out about best practice for the use of generative AI with some aspects of cultural heritage, but some people have had very good results for things like metadata generation. We need to keep experimenting and documenting our succeses and failures, to we can produce trusted best practice for the use of AI. Otherwise we risk contaminating our trusted information repositories.

Interdisciplinarity and Inclusion

How do interdisciplinary collaborations contribute to the development of inclusive and ethically grounded digital humanities projects?

I think interdiscplinarity is key – we need to bring together knowledge from across different experts and domains. Digital is so complex these days – gone is the time when one person could know everything about different technologies, systems, and their effects – and so we need to work together in teams with a variety of voices included, especially to make sure that there are no harms created in the use of technology, particularly in the use of AI. We can only do that if we think about things from a variety of viewpoints, rather than thinking any one of us knows all the answers. We can only build a better future if we work collaboratively, together, with different skills and bringing together different expertise.