Tell us a bit about your background

My PhD was actually in English literature, on authors writing from prison in Tudor England. However, about 11 years ago I started collaborating with the Turing’s Sebastian Ahnert (coincidentally my husband) on the application of methods from network science to the study of Tudor letters, and we ended up doing two projects together, funded by the Arts and Humanities Research Council. I have gone on since to collaborate with larger teams of data scientists and humanities scholars.

What are you working on at the moment?

I spend most of my time leading the Living with Machines project. One of the pieces of work that we’re just concluding is a history of the idea of the ‘living machine’, a kind of Victorian precursor to artificial intelligence. We developed an analysis method that finds mentions of animate (and inanimate) machines in text, and applied this to 19th century books.

How can the humanities help data science and AI?

We bring, and think about, data that helps data scientists sharpen their critical and technical skills. Humanities data is really big (because we have the whole of history to look back on); it can sometimes be exceptionally dirty (i.e. lots of errors); and it is often incomplete. It therefore provides important test cases for modelling bias and missing data – because, guess what, humanities data is not the only data that’s biased! This kind of thinking has really important ramifications when engaging with language models, for example. If we train language models on biased data, they can end up generating toxic language – we see this happening time and time again.

What’s the most surprising thing you’ve discovered with the Living with Machines project?

The breadth of objects to which the word ‘machine’ was applied in 19th century newspapers: looms, bicycles, bathing huts, prams, trains...

What is your highlight of the Living with Machines exhibition in Leeds?

I think it will be seeing the screens featuring visualisations of data worked on by the Living with Machines team. We’re used to seeing our work published in articles, but not on the walls of galleries!

And finally, when not working what can you be found doing?

Wrangling small children! I have a 4-year-old and a 7-month-old, so that keeps me pretty busy!