Silicon for Orchestra and Articial Intelligence

https://aimc2023.pubpub.org/pub/wbt654g2

This paper discusses the recent large-scale orchestral work ‘Silicon’, composed for the BBC Philharmonic. ‘Silicon’ uses and examines artificial intelligence in the context of orchestral music in three key ways, divided between its three movements: as a symbolic-generative ‘composer-like’ agent, as a tone-transfer ‘instrument- like’ agent, and as a neural synthesis ‘performer-like’ agent. These strategies are discussed alongside novel compositional ideas discovered through working with AI, and the broader question of how a performing institution, such as an orchestra, might incorporate advanced technologies into its fabric. The paper situates AI as a useful and creative tool for composing new music for orchestra, and the orchestra as a useful and creative tool for examining AI.