(a few words originally posted on Twitter on March 27, 2023)
I’ve been trying all morning to figure out what bothers me about these Mid journey-generated manuscripts without simply sounding like a Luddite, and I think I finally have it.
It’s because my interest in manuscripts is almost entirely about the humanity behind them. Who made them? Who used them and why? What happened to them after they were made? Where are they now? What did they mean in the past and what do they mean now?
A computer generated book doesn’t have any of that context. I’ve talked about the uncanny valley with regard to digitized manuscripts, and this is that, one step further. It’s one thing to digitize a manuscript in a way that elides its materiality, and a whole other thing to create manuscripts that don’t exist materially at all.
I think there are potentially interesting ways to use AI in my work. I’m interested in structure, and have been part of a project, VisColl, to develop models and software to build models of manuscripts. Could AI be used to combine structural models and digital images to create photorealistic imagery of existing manuscripts? Imagine an AI reconstruction of manuscripts cut apart and distributed by Otto Ege. Could it even generate pages that are lost as semi-realistic placeholders?
Just a few thoughts. I’m less interested in generating realistic looking manuscripts than in the potential to leverage the technology to help us understand the use and history of manuscripts that exist in the real world.
Added: If you’d like to hear me talk more about manuscripts and humanity, check out Coffee With A Codex, a weekly 30-minute program both live and posted to YouTube where I present a show-and-tell with books from the University of Pennsylvania’s premodern manuscript collections, and Inside My Favorite Manuscript, a weekly podcast I do in my own time where I talk to people who love manuscripts about manuscripts they love the most.