Keep up to date with upcoming MusAI public events, seminars and gigs – information about attending below
And get up to speed on our past MusAI events
Contact us at MusAI for information about attending future MusAI events. Please let us know when you write who you are, what your interest in the program is, and how your own work relates to music AI: thank you.
Upcoming events
Seminar 12: Work-in-Progress: Composing the Assemblage
7 June, 18:00-20:00 BST.
Artemi-Maria Gioti, Aaron Einbond, Georgina Born
Artemi-Maria Gioti and Aaron Einbond will present work-in-progress on their collaborative paper with Georgina Born, “Composing the Assemblage: Probing Aesthetic and Technical Dimensions of Artistic Creation with Machine Learning”. In our study, we address the role of ML in the composition of two new musical works for acoustic instruments and electronics through auto-ethnographic reflection on the experience. Rather than focusing on narrowly defined ML algorithms, we take into account the eclectic assemblage brought into play: from composers, performers, and listeners, to loudspeakers, microphones, and audio descriptors. Tracing the creative process of composing these works, we focus in particular on the aesthetic implications of the many nonlinear technical decisions involved in composing the assemblage, applying theoretical frameworks of material engagement and critical data studies. Our findings include a deconstructive critique of data as contingent on the decisions and material conditions involved in the “data-making” process. We also explore how engagement with the assemblage of ML agents has significant similarities—as well as important differences—with existing models of material engagement.
We welcome your discussion and critique of our text-in-progress, with will follow one week before the seminar.
Past events
Seminar 11: In Search of the Humans in Machine Listening
19 April, 20:00-22:00 UTC.
Jonathan Sterne, Mehak Sawhney and Andy Stuhl
The goal of this seminar is to discuss a paper in progress.In this paper, we catalogue a range of constructs of “the human” operating in machine listening systems, from abstract mathematical models, to cultural desires and ideals, to actual human listeners. Machine listening is machine learning (often branded as “artificial intelligence”) that either deals with sonic data, or uses data to produce sound. Machine listening includes a range of fields that are not always in conversation with one another: music information retrieval, natural language processing and speech synthesis, computational auditory scene analysis, and other areas.
Drawing on readings of scholarship and practice in machine learning, our paper is an exercise in critical taxonomy. Across the fields of machine listening, we group the vestiges of the human we encounter into four broad categories: imitative (like humans), delegated (like a servant or labourer), objectified (like a category), and aesthetic (“arty”—like an aesthetic product, practice, or judgment). In order for these humans to work within a machine learning system, they all must be operationalized in some way. But many of the most politically problematic aspects of machine listening come precisely from this need to operationalize. Conversely, many of the most valuable aspects of sonic culture are the ones that are most difficult to operationalize in a data-driven machine listening system.
Some recent scholarship in the humanities and social sciences has cast machine learning as an occasion to further advance arguments in favour of posthumanism, focusing on the alien dimensions of automated machinery. Conversely, this paper argues that “the human” is very much alive—if not always well—in machine listening systems. Thus, alongside arguments against excessively anthropomorphizing machine learning systems, we suggest scholars beware of excessive deanthropomorphization of so-called AI.
Seminar 10: Mock Tudor: Engineer-led collaboration on the Neural Network Synthesizer, 1989-1994.
29 March, 20:00-22:00 UTC.
Christopher Haworth
This work in progress talk examines the collaboration that led to the Neural Network Synthesizers used by experimental musician David Tudor in the later part of his life. The project was made possible by the donation of an analog neural network chip, the ETANN 80170NX, by Intel computers. A casualty of the downturn in the ‘second wave’ of AI development and investment at the end of the 1980s, the ETANN’s failure to find a use in vehicle control or robotics enabled one of the company’s engineers, Mark Holler, to explore a more unconventional application. The neural network synthesizers he helped build were not intended to simulate human intelligence, and nor were they primarily intended as a critique of projects that took this as their aim. Nonetheless, AI discourses may help in analysing the dynamics of the collaboration that led to their development. Although Holler and fellow engineers Forrest Warthman and Mark Thornson credited Tudor with steering the project as a ‘guiding light’, I argue that he took a more passive role. Less an artistic director, it was as though his oeuvre and performance practice functioned as a corpus, ‘training’ the design of a new, Tudor-esque, machine. The talk assesses this dynamic in relation to existing theories of relayed creativity in the arts, asking whether the questions Holler et al’s work posed about the role of engineers in experimental music may have restated those once asked by Tudor about the role of performers.
Seminar 9: The Politics of AI and the Politics of Technology
Wed March 8, 20:00-22:30 UTC
In this round table we want, collectively, to get to the heart of some urgent yet classic and related questions: What are the politics of AI? How do they relate to how the politics of technology have been conceptualised by STS, media studies and music/sound studies? And how do the politics of AI relate to critical research and to critique of the kind MusAI aims to cultivate? To throw light we have invited a series of speakers to present provocative 5-7 minute statements that address their take on these questions. Our aim is in part collectively didactic: to make links between the AI literature and earlier theorisations of the politics-technology relation; to make us aware of distinctive traditions that have addressed these questions, and not to lose sight of links to earlier approaches; and (of course) to reinvigorate those older traditions by taking the measure of their relevance and adequacy when faced with AI. We aim for lively discussion and participation! Please join us for this tune-up as a preliminary to the November MusAI conference, where these issues will haunt our concerns .
Seminar 8: On the cultural mediation of music AI: challenges to ‘strong Foucaultian’ readings
24 January, 20:00-22:00 UTC.
Gustavo Ferreira and Georgina Born
Gustavo Ferreira and Georgina will present two short papers for discussion that address the relationship between AI and culture with reference to recent studies of music streaming platforms and music recommender systems (MRS). Both papers engage with the advances and limitations of some recent writing particularly around the subjectification of platform and MRS users, presenting alternative framings that draw on classic arguments from STS and cultural studies, and particularly on the work of the late great Latin American theorist Jesús Martín-Barbero. On this basis both papers present alternative framings to the existing literature. Gustavo’s pre-submission paper focuses on accounts of distinctive live music scenes under strict COVID-19 movement restrictions that engaged with the affordances of streaming platforms mediated by shifting notions of liveness, locality and authenticity. Georgina’s short paper sets out four articulations between AI and (music) culture, an exercise in conceptual clarification, while also pointing to a major conceptual question posed to a classic STS model by the material nature of AI sociotechnical assemblages. She raises: does the materiality of AI assemblages mean that what she identifies as ‘strong Foucaultian’ theorisations of users are more apt than for earlier generations of music technologies?
MusAI Conference School of Advanced Study, UCL, 28-29 Nov.
Performing Critical AI I: feedback, noise, corpus, code
Sunday 27 November | 2pm
Cafe Oto, 18-22 Ashwin Street, E8 3DL
£12 | £10 in advance. Buy tickets
Anna Xambo | P.A. Tremblay and Owen Green | Feedback Cell feat. Ollie Bown
With the explosion in music technologies offering ‘artificial intelligence’, artists and musicians are exploring original and meaningful ways to adapt them to creative ends — often in ways that critique their underlying assumptions. Computational systems will be used to explore themes of agency and performative creativity, and to find new ways to control spatialisation, compose algorithmic patterns, and respond to bodily gesture. The performances will be followed by open Q & A and discussion with the artists.
Performing Critical AI II: body, space, action, agency
Tuesday 29 November | Doors: 8pm | First Act: 8:30pm
IKLECTIK, Old Paradise Yard, SE1 7LG
FREE ENTRY
Xenia Pestova | Artemi-Maria Gioti | Maxime Echardour | Aaron Einbond | Christopher Haworth
Prepared piano, handmade percussion, new compositions, and electronic improvisations situate AI with the listener in a unique 3D sound environment. With the explosion in music technologies offering artificial intelligence, artists and musicians are exploring original and meaningful ways to adapt them to creative ends — often in ways that critique their underlying assumptions. Live performances explore themes of agency and performative creativity to find new ways to control spatial sound, compose algorithmic patterns, and respond to bodily gesture. The performances will be followed by open Q & A and discussion with the artists.
Seminar 6: Critically composing with AI – work in progress
October 12, 2022. Artemi-Maria Gioti & Aaron Einbond
Artemi-Maria Gioti and Aaron Einbond presented work on their project “Permeable Interdisciplinary: Algorithmic Composition, Subverted (WP3c)”. Artemi presented her composition Bias II for piano and electronics, to be premiered in November 2022 and Aaron presented his composition Prestidigitation for percussion and 3D electronics, premiered in September. In our discussion of both works we explored themes of aesthetics, action, affordance, and agency.
Seminar 5: Redesigning recommendation
September 7, 2022. Georgina Born, Fernando Diaz, Gustavo Ferreira & Andres Ferraro
Recommender systems (RS) have become the dominant means of curating cultural content, including music. In this seminar we invited discussion of our interdisciplinary (CS + SSH) research on alternative designs for RS informed by normative principles oriented to the public good (Andrejevic 2013; Born 2006, 2018). To identify useful normative principles to guide RS design, in the first phase of our work we examined those principles underpinning public service media (PSM) systems as well as critical debates over fairness, bias and discrimination – to assess if they could be productive for redesigning RS.
Seminar 4: AI and commercial music production
June 8, 2022. Oliver Bown & Eric. A Drott
Small startups and big tech companies are now very active in pursuing commercial opportunities in the application of AI to music. In this seminar we scanned and analysed this activity, considering the emerging business cases, issues of copyright, ethics, attribution and authorship, the sociotechnical challenges of developing transformative technologies that catch on, and the possible impacts of such technologies.
Seminar 3: Ethnography of algorithms and platforms
May 11, 2022. Nick Seaver & Darci Sprengel
This seminar focused on the methodological challenges of operationalizing and studying “algorithms” and “platforms” ethnographically. It was led by Nick Seaver and Darci Sprengel, drawing on his completed fieldwork with developers of music recommender systems in the US and her ongoing work in the Arabic music industry and local music scenes in Egypt. The selected readings explored the challenges of rendering abstract technical objects and processes suitable for ethnographic study, as well as considering what kinds of questions ethnographic methods are (and are not) suitable to answer. Nick’s work was put in dialogue with Georgina Born’s earlier ethnography of a (music) AI research culture (the 1996 reading below) to highlight another, complementary angle on doing ethnography of this kind and what it can address.
Seminar 2: Software studies
April 13, 2022. Christopher Haworth & Aaron Einbond, with respondent Tobias Blanke
In this seminar we explored some of the software studies literature, looking at how its more formal orientation can complement the ethnographic approaches of the MusAI subprojects. The seminar was guided by questions like: how should we think the textuality of code, and what interpretive methods should we bring to its analysis? How do we analyse collaboration using version control systems like GitHub? What kinds of empiricism has software studies typically entailed, and could it entail in the future? We will discuss potential connections to Aaron Einbond’s and Artemi Maria Gioti’s work-in-progress on Permeable Interdisciplinary: Algorithmic composition, subverted (WP3c).
Seminar 1: Interdisciplinarity
March 9, 2022. Georgina Born, Fernando Diaz, Gustavo Ferreira & Andres Ferraro, with respondent Lucy Suchman
The opening MusAI seminar focused on interdisciplinarity in the practice of Artificial Intelligence in music and, reflexively, in our own research practice. The discussion covered three paradigms: 1) Barry and Born’s ‘modes’ and ‘logics’ of interdisciplinarity (Barry and Born 2008, 2013); 2) Galison’s ‘trading zones’ (1997); and 3) Star’s concept of ‘boundary objects’ (Star 1989, 2010; Bowker and Star 1999). The seminar also included a short presentation on interdisciplinarity in practice by Fernando Diaz and Georgie Born stemming from their project ‘Interdisciplinary Interventions in the Design of Recommendation Systems’ (WP4c), based at Mila in Montreal, a 4-way collaboration with Gustavo Ferreira and Andres Ferraro. Central to the project are our attempts to experiment by developing deep interdisciplinary exchanges between computer science and SSH, and we will reflect on this and welcome input from the group. Lucy Suchman, from the MusAI Advisory Board, responded to the presentations with reference to her own experiences of interdisciplinarity.
MusAI 