Datapower: Prediction and Incomputability – Luciana Parisi (Duke University)
As machine learning AI has replaced symbolic logic with predictive reasoning, no longer can the means or instruments of capital be associated with teleological metaphysics and ontological totalities. Whilst some argue against the data infrastructure that reproduces the cognitive mapping of sovereign computation, others have instead defined intelligent machines as instruments of war against populations (or the living itself). This talk discusses these views in light of a profound transformation in the history of machine intelligence demarcated by the automation of learning in the fields of cybernetics and computation. From this standpoint, this talk suggests that a critique of datapower requires a challenging re-invention of the notion of instrumentality at the core of critical theories of technology. In order to shift the underlying limits of today’s polarised views about data (e.g., data appropriation and data refusal), this talk will discuss the dynamic tension between predictive intelligence and incomputability in learning data systems.

Biweekly on Tuesdays – 6pm – AH 217/218