|
Artificial Intelligence as an Archival Science| title | Artificial Intelligence as an Archival Science |
|---|
| start_date | 2026/03/20 |
|---|
| schedule | 11h |
|---|
| online | no |
|---|
| location_info | visioconférence Big Blue Button |
|---|
| summary | Mass digitization of the record of human language, history, and culture has accelerated over the past two decades, but the benefits have not been evenly distributed. The potential of AI systems to act as computational models of human perception and reasoning can inform new research questions in the humanities and social sciences, as we build better methods to understand how individuals, organizations, andsocieties process information. The humanities and social sciences, in turn, provide an impetus for artificial intelligence to realize its potential as an archival science, engaging with the human record not only to mine "raw material" for training representation and prediction models but to remediate human culture for understanding. In this talk, I describe work with my collaborators on making archives of human culture computable and accessible, including our progress on optical character recognition for manuscript and print, text reuse analysis for tracing cultural influence, models of citation and memorization, and using image generation models to analyze text. I close with some observations about the gaps in the digital record and the potential for computational methods and data infrastructure to help us imagine their repair. |
|---|
| responsibles | Bawden |
|---|
Workflow history| from state (1) | to state | comment | date |
|---|
| submitted | published | | 2026/03/18 12:00 UTC |
| |
|