PERFORMING AI. DATA-CRITIQUE AND DATA-POLITICS
Gespräch mit Wendy Hui Kyong Chun und Florian Sprenger
Wendy Hui Kyong Chun, Canada 150 Research Chair in New Media, Simon Fraser University, Vancouver. Leader of Digital Democracies Institute
Florian Sprenger, Professor for Virtual Humanities, Ruhr University Bochum, Germany. Director of Virtual Humanities Lab
Host: Martina Leeker. Theatre and Media studies. Aesthetic Theory and Practice, University Köln
Wendy Chun and Florian Sprenger discuss the technological, and epistemological conditions of Artificial Intelligence, long before its starts operating. Florian Sprenger is interested in the question, how the world comes into the computer via sensors, and how the incoming data are filtered and organized by algorithms. It becomes obvious, that AI is grounded in a virtual world-making, which doesn’t know anything about the world, but just operates self-referential data. The upcoming uncertainty of this condition is the background for techno-human co-operations in semi-automated devices, as e.g. self-driving cars. Wendy Chun talks about the pre-history of today’s automated sentiment analysis in sociological experiments in Japanese internment camps in Canada and the USA during WWII in the 1940s. She explores this racist project, which tested people under stress, looking for a social engineering of humans living together, and being governed under these conditions. What does this pre-history tells us about the aims and effects of automated sentiment analysis, applying the former models and methods for optimizing products and customers’ buying behavior? Also today’s recommendation systems are still constituted in a history and in technologies of problematic social sciences, dealing with models of segregative and racist homophily, the belief that the same sticks to the same, and eugenics for social engineering.
The talks follow up with a discussion of Wendy Chun’s notion of “algorithmic authenticity” in recommendation systems, showing a regime of forcing users into fixed identities and homophilic neighborhoods, fearing users’ intentionality and collective operations.
This episode is an important folio of reflecting about performing with AI, on stage or in nettheater, which are confronted with the algorithmic regimes of Artificial Intelligence, as well as with their social engineering. The question is, how the “drama of big data” (Wendy Chun) could be performed differently, bringing: differences, diversity, plurality, latencies (instead of fixed orders, and authentic identities), and open futures. And: Which traps should be avoided so that performing with algorithms doesn’t become just a training for algorithmic authenticity. Theatrical methods are to be explored as: using the theatrical as-if, playing around with modes and models of intentionality, or inventing and performing collective interventionist actions.
Wendy’s writings on the topic, a selection:
Florian’s writings on the topic, selection:
(2021) Autonome Autos – Die Zukunft des Verkehrs und die Dispositive der Mobilität. Transcript, Bielefeld.
Intervals of Intervention. Microdecisions and the Temporal Autonomy of Cars. In: Stine, Kyle/Volmar, Axel (2021): Media Infrastructures and the Politics of Digital Time: Essays on Hardwired Temporalities. Amsterdam, Amsterdam University Press, S. 157-176.
Microdecisions and autonomy in self-driving cars: virtual probabilities. In: AI & Society 2020.