Oliver Bown, Associate Professor in Computational Creativity and Generative Music & Art, University of New South Wales, Sydney
This project investigates the dramatic rise of commercial AI music enterprises, particularly AI music startups. It looks at the multiple value propositions being put forward by these enterprises, and the technological strategies and problems being encountered by companies that must manage the dual tasks of technological innovation and building and maintaining a viable product. The project asks: how is innovation and transformation of musical practices driven in these complex sociotechnical systems?
To address this question I examine the myriad interplays between technological and social elements within a number of AI music commercial enterprises. In each case, the company must navigate between the development of a product and the development of a user base, where both are relatively underexplored: we don’t know how well the software will do its intended job, what architectures and interface designs will be successful, what patterns of use will emerge amongst users, and what types of role AI will play in the resulting products.
In such cases companies must balance the need to engage in an exploratory search for solutions with the need to support and maintain an existing product. Following theories of innovation and creativity, the role of blind, exploratory search for new designs and use cases in these companies’ search for product-market-fit is a particular point of interest. I look at how companies take risks and respond in agile ways to user trends and new technology discoveries, but above all, at the extent to which they simply explore hunches and unformed ideas. In relation to this I look at how the technology itself pushes and pulls its developers through its own internal logic, and consider how these companies’ product innovations take an emergent form. I consider whose or what agency is involved in driving this emergence. How do companies’ narratives of value around AI’s potential to transform music combine the elements of clear product vision and space to allow blind search?
A consequent and more speculative question is the extent to which such efforts may be driving an intensifying industrialisation of music creation, one that is taking entirely new forms, and how this might be seen as part of longer term processes of technology and change. One specific focus is the way that, as technology changes, different things become possible with music listening, in particular the flexibility of the musical work.
This research has resulted in four areas of study, captured in four papers and now being synthesised into a monograph, which seeks to integrate these strands and provide a comprehensive overview of how a range of music technology initiatives interact with conceptions of music’s social and individual value.
Bown, O. (2023). Music AI’s Potential Impact: Scoping the terms of the debate about value. AI Music Creativity, 2023, Brighton, UK.
ABSTRACT: The debate around the impact of AI on creative industries such as music has become increasingly mainstream as innovations in creative AI technologies appear apace, and companies are seeking commercial opportunities in the field. Often these debates are only loosely grounded in holistic thinking about value. Using academic frameworks for understanding value in the arts, we can think more formally about AI’s impact on different arts stakeholders and practices. This paper draws on critical work on value in the arts, applied to current debates about AI’s impact on music, illustrated by recent online discussions. It concludes that equal attention needs to be paid to the intrinsic as well as the instrumental value of creative production, and that arguments around commercial and professional aspects of music need to be expanded in terms of societal benefits, and with an awareness of the social context of competitive individualism and its alternatives.
https://aimc2023.pubpub.org/pub/rwi3v7tb/release/2?readingCollection=4d9db16c
Bown, O. (2024). Blind search and flexible product visions: the sociotechnical shaping of generative music engines. AI & SOCIETY, 1-19.
ABSTRACT: Amidst the surge in AI-oriented commercial ventures, music is a site of intensive efforts to innovate. A number of companies are seeking to apply AI to music production and consumption, and amongst them several are seeking to reinvent the music listening experience as adaptive, interactive, functional and infinitely generative. These are bold objectives, having no clear roadmap for what designs, technologies and use cases, if any, will be successful. Thus each company relies on speculative product visions. Through four case studies of such companies, I consider how product visions must carefully provide a clear plan for developers and investors, whilst also remaining open to agile user-centred product development strategies, which I discuss in terms of the ‘blind search’ nature of innovation. I suggest that innovation in this area needs to be understood in terms of technological emergence, which is neither technologically determinist nor driven entirely by the visions of founders, but through a complex of interacting forces. I also consider, through these cases, how, through the accumulation of residual value, all such start-up work risks being exapted for more familiar extractive capitalist agendas under the general process that Doctorow calls “enshittification”. Lastly, I consider a number of other more specific ways in which these projects, if their growth is achieved, could influence music culture more broadly.
Click to access s00146-024-01862-x.pdf
Bown, O. (2025). From genies performing magic to sages imparting wisdom: a value-centred survey of music AI user interfaces, creative affordances and artist objectives. Journal of New Music Research, 1-14.
ABSTRACT: There are, at present, multiple fronts of debate about the ethics of Commercial Creative AI (ComCAI) product development. The issues of copyright and the right to know the AI provenance of work are dominant and well-defined. Meanwhile, a significant area of debate is trickier to define and analyse, relating to ComCAI’s impact on the intrinsic values and virtues of human creative endeavour. In this paper, I analyse this impact by surveying the emerging landscape of user interfaces (UIs) in the domain of commercial music AI products. I do this through a value-centred design lens, considering how it is possible to articulate intrinsic values in the arts. Specifically, I select two artistic objectives: the need for creative artists to create works that are differentiated from others; the need for deep cognitive creative engagement with the work being created. I ask how ComCAI UIs can enable these artistic objectives, first through a survey of UI types and their combination in products, second through user feature requests. This design analysis tentatively suggests that those ComCAI tools promising to support artists increasingly need to provide greater control, in order to satisfy their creative users, and this is focussed towards either simply allowing post-generation editing, or improving generative control, which is in tension with the opaque nature of the generative systems. I propose that a value-centred design approach that highlights values such as differentiation and creative engagement helps to identify under-explored design spaces which have the potential to make better music AI experiences.
https://www.tandfonline.com/doi/full/10.1080/09298215.2024.2442360
Bown, O. (In review). The Song Remains the Same and the Beat Goes On: Prospecting the Rise of DATaPoP (Deciding Artistic Things at Point of Playback).
ABSTRACT: A recent wave of investment in music AI startup companies appears to herald significant transformations in music culture. Amongst many possible impacts, one of the most profound would be the dislodging of the fixed-media recorded work (hereafter the “track”) as the dominant musical object, upsetting vast frameworks for musical exchange, and of our collective understanding of what music is. Several companies have made this transformation their direct mission, for example by envisioning forms of musical experience based on generative streams. Additionally, other companies are pushing “source separation” technologies, those that allow tracks to be “unmixed”. Such technologies break the inviolability of the track: giving audiences just as ready access to the components of the music as artists (already a built-in feature of some DJ software). Several other technological and cultural developments may compound the forces driving a move away from fixed media from within the AI sphere and beyond. For example, the long-term trajectory of spatial audio formats, most notably Dolby Atmos, points to artists handing over greater flexibility of mix elements at the point of playback. This paper seeks to bring together evidence from a wide range of sources, including studies of music AI start-ups, public debates, a review of available technologies, and the author’s own creative practice, to inform an understanding of such potential transitions in music. It speculates on the musical objects and roles that can be seen emerging from these technological changes.