Much like technological assemblages before it, what is commonly understood as AI is embedded in what Sheila Jasanoff and Sang-Hyun Kim call ’sociotechnical imaginaries‘ – a material and social construction of desirable futures and common sense. As such, AI is put in place in the broadest way possible, to signal progress, futurity, security, and so on. Meanwhile, our shared technological realities seem to be more about anxiety watching and doom scrolling, marking a gap in the common held public belief and private affects. While it is true and necessary to – as Timnit Gebru and others have done – demystify certain paradigms of AI in their historical problematics, their underlying colonial legacies and problematic forms of capitalist extraction, the talk argues that within AI environments, there is no fundamental ‚truth‘ to be uncovered. Instead, AI coexists alongside and not despite its material realities of extraction, enslavement, and ecocide, while it also continues to create affective disjuncts and self-alienation. Resignifying AI as Ancestral Immediacies, my work delves into the promises of AI to unpack what they may entail, and to argue that now, more than ever, art and speculation are critical tools to respond to AI’s seeming dialectics.

Sara Morais dos Santos Bruss is a culture and media researcher and curator, as well as a board member of diffrakt. Center for Theoretical Periphery in Berlin. She works at the intersections of feminist and anti-colonial art, political practices, digital technologies and narratives of (human and non-human) subjectivity. Recently published: “Queere KI. Virtuality and myths of un/uniqueness”, in: digital:gender – de:mapping affect. A speculative cartography, edited by Julia Bee, Irina Gradinari, Katrin Köppert. Spector: 2025