Throughout academic and industry discourse within the music industries, copyright tends to feature prominently. Several have even defined the music industries categorically as being ‘copyright industries’ (Laing 2004, Wikstrom 2009), due to the central position copyright plays within the commodification of music.
Copyright has played a central part in almost all modes of music consumption – it has allowed musicians and songwriters to reproduce and sell their output on a mass scale, whether this takes the form of physical copies such as vinyl and CDs, or more abstract formats such as downloads and streams. Copyright law has evolved as technology has changed with regards to music consumption (Laing 2004), but can it grow with technological change that affects the production side of the music industry chain of production?
The most recent advancements in songwriting technology have revolved around machine learning, and Artificial Intelligence. In particular, songs have recently been written by FlowMachines, a program developed by Sony. Both songs are created by analysing a database of music and replicating stylistic signifiers to compose similar music (Goldhill 2016).
I believe this process to open considerable questions surrounding copyright legislation regarding machine-written music. The obvious answer to these questions would be to treat the database of music in a similar way to song samples in music, but on a considerably larger scale (dependent on the size of the database) (Sannie 2017). However, I believe there to be counter-arguments to this notion.
In essence, the software or machine composer is analysing a collection of music, and using this as a stylistic guide to create new music, a process I believe comparable to a human composer, who will base his or her own compositions on a specific sound, influenced by genre, musical memories and music listened to recently.
Of course, different parties will have different opinions. The creators of the software, the composers of the database music (however many there are) and the user of the software all have claims to the various copyrights attached to the resultant output. As ‘machine music’ becomes more popular (Sony have tentative plans to release an entire album of AI-generated music in 2017 (Goldhill 2016)), copyright law will once again have to adapt to the technological landscape of the time.
- Goldhill, O. (2017). The first pop song ever written by artificial intelligence is pretty good, actually. [online] Quartz. Available at: https://qz.com/790523/daddys-car-the-first-song-ever-written-by-artificial-intelligence-is-actually-pretty-good/ [Accessed 17 Feb 2017].
- Laing, D. (2004). Copyright, Politics and the International Music Industry. In: S. Frith and L. Marshall, ed., Music and Copyright, 1st Hoboken: Taylor & Francis.
- Sannie, M. (2017). Music Publishing and Disruptive Technologies.
- Wikstrom, P. (2009). The Music Industry. 1st Cambridge: Polity Press.