Everybody’s Song Making

Abstract

Since the late 1990s, Artificial Intelligence (AI) has written songs that surpassed human composers in reproducing recognizable styles. In 2016, London’s West End premiered the world’s first computer-generated musical, Beyond the Fence, based on digital technology to detect ‘hit’ musical theatre songs. As AI quickly learns and upgrades their song-writing skills, humans are developing new ways of working with AI to expand the forms of song making for homo sapiens. In our post-phonograph era of information technology, it is no longer possible for the human song maker to write, produce, perform and distribute songs without any AI assistance. Recent scholarship on algorithm bias and AI domination, however, reminds us the hazard of AI-dependent human society and the urgency of consciously working against AI. The human public DIY song making has thus contributed to a larger political discourse that resists AI and the forces behind it— technology-driven global capitalist and consumerist infrastructures. Revolutionary apps such as Vocaloid and Ocarina have made anti-AI song making attempts accessible to the public, yet are precariously entangled in global data-processing networks and commercial systems. This article explores the social dimensions of human DIY song making in the age of Artificial Intelligence. It argues that the DIY song making creates distinct virtual communities and engenders new soundscapes that not only defy past normative algorithms and patterns, but uplift under-represented voices and desist from data-centric cultural production endeavours.

Notes

1 Music samples are accessible at https://aiva.ai.

2 See ‘FlowComposer: composing with AI’ at https://bit.ly/2FnwfFb

3 See ampermusic.com.

4 See ‘IBM Watson Beat’ at https://ibm.co/2PY0Xru

5 See ‘Have yourself an AI Christmas’ at http://bit.ly/2Ws2yZt