We are a collective of researchers using humanistic tools to investigate artificial intelligence. Our objective is to use writing, a technology with its own constraint and histories, to excavate AI.

We begin by exploring the link between AI and writing. Writing runs through the history of AI: the code, the dataset, the token, the prompt, the output, and in the attempts to formalize reasoning and intelligence. Even the mathematics and formal logics of AI mirror elements of writing. What is it about AI and writing that make the two technologies so connected? 

This course is for students who want to research AI, and who sense that the most useful preparation for that is to listen to what philosophers and storytellers in conversation with mathematicians and computer scientists have to say.

The history of writing will expose the older assumptions and formal systems sedimented into all the shiny bells and whistles touted in today’s discourse of AI optimism. But AI, in turn, may end up making writing unrecognizably strange. Where machines can generate writing at scale, lots of things we take for granted about the nature and value of writing will need to be re-evaluated. Research Question: How do humanists undertake AI research? How can students be taught to do humanistic AI research. 

Imagine this course as a room full of objects that don’t belong together — a divination sand drawing from Mali, a tokenizer, an infinite library, a Parisian arcade, a perceptron — and the task is to let them speak to each other and tell us something about generative AI beyond the version that Silicon Valley sells and that the headlines scream about. We are seeking out the stranger, older, darker, more consequential thing underneath.