Motivations for using Artificial Intelligence in the popular music composition and production process.
Geelen, T. van
MetadataShow full item record
While there has been significant academic interest in AI-systems for musical composition, adoption of such systems by practitioners has been slow. Concurrently, Pearce, Meredith and Wiggins argue that research into AI is slowing because of a failure to develop and adopt an appropriate methodology. PMW proffer a solution in through better categorisation, which Sven Deserno proves hasn’t gained significant traction. Issues surrounding artificial intelligence in music are further complicated by the ambiguous nature of the term, and it’s widespread and often erroneous use in media. This thesis aims to answer questions surrounding this slow adoption of AI-systems for musical composition, the apparent hesitance towards embracing them for musically creative purposes, and practical aspects that might limit their use for composition of commercial music. It does so by juxtaposing Calestous Juma’s theory on fear of adoption of new technologies, early historica and modern-day approaches to algorithmic thinking in music and creative thinking in music composition against experiences of makers and users of AI-systems for musical composition. I will focus specifically on the music of popular singer-songwriter and AI-enthusiast Taryn Southern and AI-researcher and composer David Cope and critical analysis of said systems. I argue that the slow adoption is caused by more than mere financial incentives, as many innovations have, but rather a specific type of fear of letting new technology invade the creative process that practitioners might identify as strictly human in nature, combined with the difficulty of effectively evaluating the usefulness of AI-systems, and a general lack of affordances regarding AI in musical composition. In 2000, David Cope mentions in his final chapter of The Algorithmic Composer: “what rational composer would refuse to listen to an algorithmic attempt to solve a problem”. This thesis offers a beginning of an answer to Cope’s hopeful or woeful prediction.