Noisy Chatbots, Silent Archives, and the Illusion of Interoperability

Archival collections are increasingly being digitized; however the ease of access belies the multiple processes that result in inherently partial and selective collections online. Furthermore, infrastructures holding in place digitized materials are built based on Western worldviews and imperial legacies that mediate and condition our understanding of materials. Striving for interoperability without being aware of or interrogating omissions, power structures, and distortions underlying datasets we work with results in projects that perpetuate dominant voices and in scholarship that reinscribes and reifies archival silences into the digital realm. Simply put, we end up training our systems with vast swaths of world memory absent and make what is already visible hyper-visible. In my talk, I will discuss focusing not only on what is there, but mainly how to seek out