Matteo Pasquinelli and Arif Kornweitz contribute an essay to the catalogue of the Music Biennale in Venice.

The essay looks to composer Iannis Xenakis’ work in representing music as a multidimensional entity in polytopes in order to analyze how AI models process and generate music. The evolution of music representation from traditional sheet music to digital formats is traced and linked to the rise of models such as MusicLM. The concept of multidimensional spaces in AI is explored, emphasizing the use of pattern recognition and generation. The essay notes that AI models like MusicLM have automated techniques of parameterization and spatialization, concepts once envisioned by composers like Xenakis. Pasquinelli and Kornweitz raise questions about what AI can and cannot replace in music, particularly the artistic and political possibilities of expression that may be affected by automation.

Download PDF