AI and the Music Industry - An argument of deskilling
Me and my colleague are preparing a presentation for an academic conference, with a topic specifically focused on the role of generative technologies in the field of music; specifically the Music Industry.
The following is practically the summary of our argument; I wanted to share it with you kind people, hoping to spark a discussion on the topic. Do you believe we have pinpointed the "principal contradiction" of the issue? What are your thoughts on the topic?
Disclaimer: Apologies for the rigid language present in the summary. It is a machine-translated version, translated for this forum using ChatGPT*
Conclusion: The Fundamental Contradiction
The fundamental contradiction is therefore not AI’s incapacity to perform creative labour, but the restructuring of value-producing pathways in cultural sectors that historically remained only formally subsumed.
Generative AI does not simply bypass creative labour: it instead thoroughly recentralizes the entire commodity chain towards the hierarchy of valorisation by replacing labour-dependent production with model inference and proprietary infrastructure.
Rather than pushing these sectors toward real subsumption through standardized labour processes, platform AI threatens to collapse the heterogeneous "borders" of the industry: amateur scenes, semi-professional markets, and peripheral circuits of skill formation - from local groupings to individual musicians-as-consumers - that previously reproduced the conditions for commodifiable talent.
For this strata, the crisis is therefore not one of creativity (which alludes to a romanticized notion of art) but of reproduction: participation in the algorithmic music economy becomes economically irrational as generative content saturates streaming platforms; not because value disappears, but because platform capital captures that value directly.
In attempting to “optimize” creative labour through automation, platform capital does not undermine its own markets so much as it cannibalizes those fractions of cultural capital that depended on labour-rich, decentralized ecosystems of production. The historically significant reservoirs of producers and consumers, such as the DIY/indie scene (notorious for its "authenticity"-driven marketability and branding), amateur education, hardware and software markets, eventually become expendable once models can be trained on vast archives of existing music, and can satisfy demand at negligible marginal cost.
What thus, at first glance, appears as capital’s self-destruction is, in fact, simoultaneously also its own consolidation; albeit of a different fraction of capital. Infrastructure- and compute-based platforms that no longer need continuous input from living creative labour to extract surplus are centralized in this emerging hierarchy of valorisation. The resulting modulation in the mode of production is presented ideologically by many (including Suno's CEO) as “democratization,” but it actually entails the opposite: a radically centralized and rentier form of cultural production in which individuals perform deskilled prompting isolated from collective institutions of distribution, and employment.
This marks not a departure, but a development; a smoothening-out of the "music factories" of the past, such as Motown and the Brill Building, which were fundamentally centered on the highly-skilled (and thus harder to repress) strata of the session musician. By removing the problematic element and replacing it with the highly precarious prompter, the fordist logic of Motown effectively begins to undue itself in a dialectical fashion.
Precisely, this new "prompter" is in fact not a decentralized producer; they are a tenant of a proprietary infrastructure whose outputs re-enter the platform’s data pipelines. The prompter plays the role of both a producer of the output and the consumer of the very data-grid that fuels it; a modulation of the pre-existing forms of musical labour, which in itself is not dissolved. Instead, the session musician of old is eliminated and systematically deskilled to lower the relative level of labour autonomy within the process of valorisation.
Platforms such as SUNO operate precisely on this assumption: that they can replace existing cultural-industrial structures, depicted as inefficient or insufficiently “democratic,” with model-driven infrastructures under their exclusive control. OpenAI and other American tech firms pursue the same strategy: to constitute themselves as indispensable infrastructures of future economic life, even in the absence of demonstrable productivity gains.
In line with historical argumentations of Malm and his magisterial study of the steam machine, their wager is speculative and infrastructural rather than productive. It posits the fact that annihilating existing circuits of valorisation will allow them to rebuild new ones organized entirely around frameworks better suited to the process of subsumption and valorisation; just how in Malm's case, oil frameworks replaced steam. In our case, this includes a constant computing of rents, idealistic enlargement of data enclosures, and creation of constant, mass-level platform dependency.
In this sense, the principal antagonism lies not simply between capital and creative labour, as posited by romantic rejections of technological advancements in general and generative audio technology in particular, but between the now risen, yet precariously positioned, internally contradictuous and inflated forms of platform capital - which thrives on enclosure, automation, and infrastructural dependency - and the relatively archaic, well cemented, yet declining culture industry capital, which relied (and relies) on diverse, labour-intensive ecosystems of production and reproduction, leaving much to be further subsumed.
In the case of the music industry, generative AI thus represents not the disappearance of valorisation in the cultural industries and the downfall of the capital A Artist, but its monopolization by a new hegemonic fraction of capital seeking to reorganize creativity around proprietary infrastructures and automated reproduction better suited to frameworks of valorisation.
mute_compulsion in comradeship
AI and the Music Industry - An argument of deskilling
Me and my colleague are preparing a presentation for an academic conference, with a topic specifically focused on the role of generative technologies in the field of music; specifically the Music Industry.
The following is practically the summary of our argument; I wanted to share it with you kind people, hoping to spark a discussion on the topic. Do you believe we have pinpointed the "principal contradiction" of the issue? What are your thoughts on the topic?
Conclusion: The Fundamental Contradiction
The fundamental contradiction is therefore not AI’s incapacity to perform creative labour, but the restructuring of value-producing pathways in cultural sectors that historically remained only formally subsumed.
Generative AI does not simply bypass creative labour: it instead thoroughly recentralizes the entire commodity chain towards the hierarchy of valorisation by replacing labour-dependent production with model inference and proprietary infrastructure.
Rather than pushing these sectors toward real subsumption through standardized labour processes, platform AI threatens to collapse the heterogeneous "borders" of the industry: amateur scenes, semi-professional markets, and peripheral circuits of skill formation - from local groupings to individual musicians-as-consumers - that previously reproduced the conditions for commodifiable talent.
For this strata, the crisis is therefore not one of creativity (which alludes to a romanticized notion of art) but of reproduction: participation in the algorithmic music economy becomes economically irrational as generative content saturates streaming platforms; not because value disappears, but because platform capital captures that value directly.
In attempting to “optimize” creative labour through automation, platform capital does not undermine its own markets so much as it cannibalizes those fractions of cultural capital that depended on labour-rich, decentralized ecosystems of production. The historically significant reservoirs of producers and consumers, such as the DIY/indie scene (notorious for its "authenticity"-driven marketability and branding), amateur education, hardware and software markets, eventually become expendable once models can be trained on vast archives of existing music, and can satisfy demand at negligible marginal cost.
What thus, at first glance, appears as capital’s self-destruction is, in fact, simoultaneously also its own consolidation; albeit of a different fraction of capital. Infrastructure- and compute-based platforms that no longer need continuous input from living creative labour to extract surplus are centralized in this emerging hierarchy of valorisation. The resulting modulation in the mode of production is presented ideologically by many (including Suno's CEO) as “democratization,” but it actually entails the opposite: a radically centralized and rentier form of cultural production in which individuals perform deskilled prompting isolated from collective institutions of distribution, and employment.
This marks not a departure, but a development; a smoothening-out of the "music factories" of the past, such as Motown and the Brill Building, which were fundamentally centered on the highly-skilled (and thus harder to repress) strata of the session musician. By removing the problematic element and replacing it with the highly precarious prompter, the fordist logic of Motown effectively begins to undue itself in a dialectical fashion.
Precisely, this new "prompter" is in fact not a decentralized producer; they are a tenant of a proprietary infrastructure whose outputs re-enter the platform’s data pipelines. The prompter plays the role of both a producer of the output and the consumer of the very data-grid that fuels it; a modulation of the pre-existing forms of musical labour, which in itself is not dissolved. Instead, the session musician of old is eliminated and systematically deskilled to lower the relative level of labour autonomy within the process of valorisation.
Platforms such as SUNO operate precisely on this assumption: that they can replace existing cultural-industrial structures, depicted as inefficient or insufficiently “democratic,” with model-driven infrastructures under their exclusive control. OpenAI and other American tech firms pursue the same strategy: to constitute themselves as indispensable infrastructures of future economic life, even in the absence of demonstrable productivity gains.
In line with historical argumentations of Malm and his magisterial study of the steam machine, their wager is speculative and infrastructural rather than productive. It posits the fact that annihilating existing circuits of valorisation will allow them to rebuild new ones organized entirely around frameworks better suited to the process of subsumption and valorisation; just how in Malm's case, oil frameworks replaced steam. In our case, this includes a constant computing of rents, idealistic enlargement of data enclosures, and creation of constant, mass-level platform dependency.
In this sense, the principal antagonism lies not simply between capital and creative labour, as posited by romantic rejections of technological advancements in general and generative audio technology in particular, but between the now risen, yet precariously positioned, internally contradictuous and inflated forms of platform capital - which thrives on enclosure, automation, and infrastructural dependency - and the relatively archaic, well cemented, yet declining culture industry capital, which relied (and relies) on diverse, labour-intensive ecosystems of production and reproduction, leaving much to be further subsumed.
In the case of the music industry, generative AI thus represents not the disappearance of valorisation in the cultural industries and the downfall of the capital A Artist, but its monopolization by a new hegemonic fraction of capital seeking to reorganize creativity around proprietary infrastructures and automated reproduction better suited to frameworks of valorisation.
sounds like a chapter of capital haha
as in - too derivative? :)