Description
This paper reflects on what, and whose, geopolitical futures are being imagined, enacted, and foreclosed in projects to advance, ‘responsible’ artificial intelligence (RAI). These reflections are grounded in the unease we experienced during a research council-funded project to help museums with colonial collections reflect on how to use RAI. We use this funding experience as an autoethnographic case study of the costs of how AI futures are epistemologically validated and realized by institutions.
Museums increasingly use AI—e.g., collections management and interactive visitor experiences—but such uses are particularly fraught given that UK museums are sites of ongoing colonial power (Turner and Tidy, 2020), and AI ‘depends on and was made possible’ by global colonial structures of knowledge, power, and violence (Adams, 2021; Tacheva & Ramasubramanian 2023). Our experience highlighted numerous anxieties surrounding decolonization and the fetish of positive AI futures, and we argue such anxieties and fantasies are a central part of the infrastructures that maintain the coloniality of present-day museums.
Being expected to pursue ‘positive AI futures’ in this constellation of coloniality compels us to ask: What kinds of geopolitical futures are being invested in through such projects and what might disinvestment look like? How can we imagine the decolonized museum in a way that rejects the compulsion to integrate AI and the futures such compulsion upholds? And, who is ‘we’ in light of the geopolitics of research funding structures that enable such projects?
