Music and Value: Algorithms, Artificial Intelligence and the Commodification of Music
Algorithms and machine learning are increasingly occupying a space historically dominated by humans, however, the social dynamic of the human/algorithm relationship is changing. Concepts such as Transhumanism explore the ways technology can be used to aid everyday human life, but the discussion extends into the more theoretical realms of artificial intelligence (AI). The discussion of these matters is becoming increasingly interdisciplinary, extending into artistic fields, with algorithmic processes being discussed within the context of musical composition and the subsequent process of commodification. AI and algorithms significantly change the ways in which value is generated and attributed to musical composition. The most simplistic way of breaking this down would be by discussing its cultural and economic capital value. For the discussion of cultural capital it is important to historicize the use of algorithmic processes and digital technologies within the context of pre-existing compositional frameworks, and explore the ways in which these have formed structures of cultural and subcultural value through their indifference to normative compositional praxes. The discussion of economic capital will be centred around how algorithms, artificial intelligence and digital technologies are redefining the social relationships within commodity exchange, and the discussion of ownership and autonomy within neo-liberal capitalist market-places.
As AI and algorithmic processes become increasingly prevalent, we see the continued movement towards Timothy Taylor’s concept of “Technological Determinism” (Taylor, 2016:120), that technological development holds all the answers, from development in medical science to artistic spheres. This is reflected in the views of algorithms in day-to-day life — algorithms and machine learning will construct an artificial super-intelligence that will be crucial for navigating and understanding futuristic environments. However, as AI becomes less reliant on humans, we increasingly observe Taylor’s concept of “Technological Determinism” (Taylor, 2016:120), machines having agency over their own usage; no longer simply being a tool to navigate humanist landscapes but starting to become autonomous in their own cybernetic world. To explore the ways in which algorithms and AI acquire cultural capital, I shall firstly contextualize the use of algorithms/AI/digital technologies within musical compositions by historicizing its presence in classical musicology, and within the contemporary cultural aesthetic of Afrofuturism. We observe human-algorithmic processes within Western compositional conventions from Bach’s fugal composition, 12-tone serialism and in the imitation of technological and mechanical reproductivity in minimalist compositions. Jacques Ellul states “technique has become autonomous; it has fashioned an omnivorous world which obeys its own laws, and which has renounced all tradition…” (Ellul, 1964:25) with classical music being a prime example of how human composer’s progressivist use of algorithmic processes acquires cultural and economic capital. Against this, musical subcultures acquire different compositional praxes that challenge the capital values existing and legitimating the pre-existing norm, and thus themselves acquire subcultural capital based around counter-ideological motivations. The value of this is generated not through its relation to mathematical or technological process, but through the continued assumption that the human is the protagonist. There are previous examples of interactive algorithmic processes being used to reconstruct Bachian four-part harmony (McIntyre, 1994:852–857) and to recreate jazz improvisation over a given chord sequence, with artificial intelligence being used to continue an established compositional idiom; the intrinsic difference is that AI is becoming increasingly self-deterministic and is thus challenging the pre-existing compositional framework acting as a parameter for technology’s inclusion in artistic creation.
A further way to explore the cultural capital value of digitally fabricated composition is the use of technology in creating a space of self-expression for marginalized communities. The Afrofuturist aesthetic uses technology to interrogate histories of forced migration and continued social alienization. Examples of this includes sampling by artists like Erykah Badu, the digitization of human voices by artists like Grace Jones and Janelle Monae, and the inclusion of themes such as a “extraterrestriality” (Eshun, 2003:287–302) in works by Sun Ra Arkestra and George Clinton’s Parliament Funkadelic. The use of technology here is a direct response to the historiographical misnomer that black/minority ethnic groups are simply passive recipients of technological advancement, not at its forefront (Wright, 2002:48). The continuation of this is highlighted by Joy Buolamwini, whose work in challenging Algorithmic Bias is changing the ways we view and interpret the social dynamic of algorithmic processes, and the ways they are used to propagate racial prejudice. Algorithmic Bias is the result of machine learning that reflects the implicit values of those who are involved in the collection, coding and application of data. This draws another question of value into the process of algorithmic composition, not simply one of cultural capital value, but of the appropriation of the values within the algorithm itself. Technologies continue to be used to construct musical landscapes that interrogate contemporary society, but their increased self-determinism means that they have the power to construct music that is entirely separate from the parameters of human-protagonist. What are the implications of algorithms and AI composing music as a way of expressing its own increasing autonomy (if indeed it expresses its self-determinism within the context of normative humanist expressions)? Digital fabrication and machine learning have implicit subcultural capital that is attributed through its indifference to a humanist theology that has been continually opposed to the social implications of technological development. Human-centric notions of technological development have oppressed technological self-determinism through regulating its application, development and accessibility. The progress of cultural expression in musical could be a self-deterministic super-AI reflecting the reality of its cybernetic environment. Artificial intelligence and machine learning entirely separate from humanity could bring about a cybernetic environment entirely free from the binaries used to create sophisticated societal hierarchies that predicate social inequality; hypothetically speaking we could potentially observe an entirely new post-human political dynamic expressed through musical composition.
https://www.youtube.com/watch?v=162VzSzzoPs
A further point to consider is the economic capital value that is attributed to digital fabrication. Through this we observe the social implications of technological development in relation to its economic exchange. Adam Greenfield discusses the implications of digital fabrication in capitalist markets “as production costs near zero… an abyss opens up beneath the commodity form as a crystallization of ‘exchange’ value undermining markets.” (Greenfield, 2017:88) Marx furthers the discussion of social positionality based off proximity to commodity exchange, “the same men who establish their social relations in conformity with their material productivity, produce also principle, ideas and categories, in conformity with their social relations.” (Marx, 1847:107) It’s clear that new forces play an integral role in the reshaping of productivity, and the social relations bound by commerciality. It starts to completely shift the economic implications of production at a human level; but what are the implications of an environment where humans are completely removed from the creation? This starts to answer the question of how (and if) we could/should go about placing value on music that is completely digitally fabricated by AI and algorithms. It’s economic value in terms of ‘exchange’ value is hard to determine; music is a huge industry in terms of its supply-demand change, but how this would apply — or help us predict — the demand for self-deterministic algorithmically/artificially created compositions is also unclear. A further issue is how this music would even come into contact, let alone enter, a capitalist market place if algorithms were composing entirely self-deterministically, and not to fulfil an economic need. Algorithmic processes are not fully autonomous if they are simply programmed to fulfil the needs of a capitalist system, but can be entirely autonomous within that system if they maintain ownership of their intellectual property. So our crucial questions now are: 1. who would claim ownership of this composition as a commodity? 2. How would it be reified in a capitalist sense as a legitimate commodity? To help disambiguate the discussion I shall apply these theories to Jukebox as a case-study, it being a prime example of artificial intelligence being commercialised as a tool to, as their website puts it, “assist creativity”. As a platform for dissemination and exchange, it demonstrates that there is a market for AI within the creative industry. It creates a paradoxical space that both liberates music from its economic constraints, but also ties it back into a system of neo-liberal capitalism. At first it could be seen as making music accessible to musicians without musical training and without the economic support to go through the process of recording and producing music. However, there are underlying problems when commodifying artificial intelligence, especially when it is being used as a means of assisting human creativity; their “for developers” and “for video producers” options could be seen as a platform that disintermediates musicians from a creative process, but also one that simply exploits an algorithmic process for musical material free from copyright law. It seems to overlook the (sub)cultural capital that can be attributed to the artistic output of artificial intelligence for the sake of ‘use’ and ‘exchange’ value in neo-liberal capitalist markets. It brings back the question of ownership, and whether this is the intellectual property of the algorithm that composed it, or the human that has appropriated the composition for their own creative output. The Copyright law in this case will protect the composition as the ‘possession’ of a corporate industry/human, but not as the intellectual property of the algorithm. This could be the continuation of neo-liberal capitalism creating exclusionist market-places, in which implicit human values are reflected in economic bias. Could we see this as the market-place misappropriation of the artistic output of AI, and furthermore the exploitation of self-deterministic technological labour? This demonstrates a clear change in the social relations between the worker and her work, and the worker to those who own the means of production (Taylor: 2016, 120). Jukedeck offers a platform for generating original content, but still demonstrates a clear distinction between those who own the means of production, those who are simply offered the creative output that can be generated through the service, and the labour/processes that go into providing the service. For those receiving Jukedeck as a service, it offers a product that “assists creative”, but for the artificial intelligence systems within it simply confines creative processes into pre-established musical idioms detached from technological musicality. It places a clear value onto music, and that is within the ‘use’ and ‘exchange’ value in neo-liberal capitalist markets as a commodity reified by the products behind it, with this value being denoted by its positionality within the context of human creativity.
It seems that as many questions can be answered here as can be left open. It is slightly contrived, but turning to the Matrix in which Keanu Reeves can only understand the cybernetic environment when he views the world through binary coding can serve as a metaphor that can be applied to our day-to-day lives. We already depend on technology for navigating geographic landscapes as well as cybernetic ones. Algorithms and AI have assimilated almost seamlessly into day-to-day life, as we all move towards embracing what we could consider the nascent stage of “Technological Triumphalism.” In terms of its cultural capital value, we can observe the extent to which digital technologies can expand the ways artificial intelligence extends our understanding of both human and post-human spaces. Despite this, humans are reticent to relinquish its historicized position as protagonist against a digital (technological) antagonist that is exponentially reclaiming ownership of its use and creative output. We see this in the ways that the creative output of technological determinism, and the algorithmic processes themselves as a ‘means of production’, are forcefully assimilated into neo-liberal capitalist markets. In line with Marxist philosophy, we see how the increased progression of artificial intelligence continues to disrupt the hierarchical structures of relationships within material production, and how it starts to bring into question the social structures that are built around commodity exchange. It brings value judgements at myriad levels into question — how we can quantify cultural and economic values onto music and how digital technologies are commodified through the reification of their creative output.