Calendar

Thanks to those of you who came along to the first session of our new reading group in critical media theory, it was a great session.

Our next meeting will be on March 4^th at 3.30pm in Copland 1.112 at the University of Westminster’s Cavendish Campus. If you’re not a Westminster student or faculty member then you’ll need to sign it at reception.

This will be a special session as we’ll be discussing a paper by Prof Eric Drott (University of Texas at Austin) on the political economy of music AI, with Prof. Drott in attendance.

An abstract of Prof. Drott’s draft paper is below and will shared with those who confirm their attendance via email (P.Rekret@westminster.ac.uk).

Eric Drott

This paper examines four conditions and a contradiction shaping the recent boom in commercial music AI (and AI more generally). The conditions include the “cheap money” regime pursued by the US Federal Reserve and other central banks following the 2008-9 financial crisis; the long-term secular decline in productivity across the global economy that AI will allegedly remedy; the growing importance that assets, assetization, and rent extraction have assumed as strategies of accumulation; and the post-pandemic reckoning with material constraints, seen in both the tightening of monetary policy and the growing recognition of the intense demands that large AI models make on a number of resources (energy, water, compute, data). The contradiction that haunts commercial music AI stems from the tension between the prerogatives of private ownership and the social character of the training data that fuels machine learning. Such data are social not simply because they embody prior cultural knowledge, or because they are typically the byproduct of digitally-mediated interactions; rather, data are social in a more profound sense, inasmuch as their utility hinges on the patterns they form with other data. Data’s value, in other words, is an emergent quality—or surplus—born of their multiple relationalities. Among other things, this contradiction hints at an alternative economy for music AI: one that sees its dependence on the sociality immanent to music and music data as an incitement for AI’s socialization, with generative and predictive systems not treated as a source of privatized riches, but as an expression of public wealth.