Archive Fever & AI

A convening on generative AI, archival practice, and found text

Saturday, November 7, 2026 @ the Parkway Theatre, Baltimore MD

This convening takes its title from Jacques Derrida's 1995 book Archive Fever. In the opening section of that text, Derrida talks about the then-emerging ubiquitousness of email. He argues that because the technical structure of an archiving technology shapes what can be archived in the first place, that if email had been around in Freud's time, with its concomitant effect of dissolving the line between public and private correspondence, Freud might never have invented psychoanalysis.

Foucault, writing a generation earlier, argued similarly, that the archive isn't a passive record of what was said, it is a structure that allows and decides what can be remembered as having been said, and thus what can be said in the future about the past, about what is "known," what the facts are.

"There is no political power without control of the archive, if not of memory."
— Jacques Derrida, Archive Fever

Archive Fever & AI brings together archival scientists, computational researchers, poets, and artists for two sessions and a public performance.

The eight people gathered for this convening are approaching this question of what generative "artificial intelligence" systems mean for archives (and thus for knowledge production, for facts, and for human choices) from different angles. Some are building the infrastructure directly, rethinking institutional access and design for a digital record, and its implications. Some are working on the problem tactically, treating AI and data systems as a new locus of control, sometimes dressed up as innovation or convenience. Others are asking what gets lost, appropriated, irrevocably altered, when highly specific and contingent human output gets passed through a system whose design works on flattening and smoothing inputs and outputs. And one is working from inside the noise itself, treating signal failure and residue as the material, the differentiator, and the innately human core of the human/machine creative dyad.

Sessions

The Archive Under Pressure, 10:00am Archival scientists and computational researchers on the methodological and ethical stakes of applying machine learning to historical collections: what's gained, what's lost, and who decides.

Found Text as Method, 1:00pm Poets and artists on using AI tools to surface archival and found voices rather than replace them, work that treats the machine's output, and its noise, as raw material for creation and investigation.

Evening Performance

An 8-channel sound installation by Victoria Keddie (Spatial Sound Union), built from archival and transmission-based material and the noise in the signal, closes out the day.

Participants

Tom Lippincott (JHU), Richard Marciano (University of Maryland), Maria Antoniak (CU Boulder), George Oates (Deep Keep), Lillian-Yvonne Bertram (University of Maryland), Eryk Salvaggio (Cambridge), Mark Shepard (University at Buffalo), and Victoria Keddie (UPenn).

Organized by

Liz Flyntz and Kyle Stine, Film and Media Studies, Johns Hopkins University. Supported by the JHU Sesquicentennial Envisioning the Future Award.

More details, including RSVP information, to follow.