CCMSAI 2026

Central to questions of co-creativity and “AI”–which we define as both generative and assistive AI models built from machine learning and other computational systems–are the ways in which different types of artificial intelligence programs impact the music workplace. The following assembles highlights from a roundtable held during the conference that explores how AI music tools reshape the creative process as well as how labor operates in these spaces. Who creates AI tools for musical applications, who controls their operations, and who gets to use them? What are the costs and consequences of AI usage, from environmental destruction to power imbalances in the music industry? We came to this panel in order to discuss these crucial questions. This discussion draws on the expertise of five professionals located across popular music: scholars who study music production, producer-scholars, and professionals whose primary work is in production.

This paper examines AI-generated music as a twenty-first-century form of propaganda, analyzing how synthetic songs shape collective identity and political belonging without top-down institutional control. Focusing on the viral AI-generated song “We Are Charlie Kirk” and its accompanying YouTube video, I trace how AI tools construct an in-group world of Christian nationalist imagery, drawing on ethnomusicological and media theory frameworks alongside historical precedents in fascist radio broadcasting, Nazi martyrdom songs, and Cold War-era country music and Hollywood film. Using reverse-prompt research on the commercial platform Suno, I show how a simple prompt can generate a fully realized martyrdom ballad, revealing the racialized and class-coded assumptions embedded in AI music generation. I further show how ironic remixes and covers of the Kirk song, rather than subverting its message, extend its circulation and reinforce in-group and out-group dynamics. A comparative case study of Iranian-produced AI Lego-animation videos targeting American audiences demonstrates how the same techniques can be weaponized from outside a political community. Together, these cases suggest that AI music generation has collapsed the cost of producing emotionally persuasive, identity-forming media, enabling a decentralized form of participatory mythmaking that parallels, and extends, twentieth-century propaganda infrastructures.

Short Concert for the Last 74s of Topsy is a generative audiovisual work that departs from the case of Topsy, an elephant electrocuted in Coney Island in 1903 and filmed by the Edison Manufacturing Company, in order to interrogate the relation between technique, memory, and the asymmetry of power. In this essay, I take the original footage as the perceptual field of a system developed in Max/MSP/Jitter, which translates the pixels of the archive into sound through the technique of sonification, within a paradigm of symbolic AI made of explicit rules and without machine learning. I discuss the idea of a music latent in the image, revealed by a non-human intelligence indifferent to affect: the mapping rules are authorial, and the algorithmic testimony is produced, not discovered. To opt for the legibility of these rules, against the opacity of the black box, is a deliberately ethical and political position, enunciated from the Global South, because it makes inspectable who decides what. Drawing on the philosophy of technology, post-phenomenology, and the critique of the image, in what follows I discuss the non-neutrality of the apparatus, the risk of aestheticizing barbarism, and the exhibition of the work as a possible spectacle. By submitting the piece to the exact seventy-four seconds of the record and by fixing in film only one among the system’s infinite possible outputs, I repeat and denounce the same gesture that turned Topsy’s death into an image.

Generative AI music systems enter a production environment in which composers are already asked to work quickly, respond to temp tracks and stylistic references, deliver modular revisions, and produce music whose commercial value may depend on functional similarity. This paper examines AI music through the linked problems of labor, attribution, and platform governance. It argues that generative systems are especially disruptive where musical work has already been standardized around speed, style replication, and low-cost synchronization. The analysis distinguishes human influence from model training, separates attribution from copyright ownership, and treats human authorship as a gradient of documented expressive contribution. It then examines the movement from post-hoc AI detection toward provenance-by-design: licensed-source metadata, generation logs, content credentials, watermarking, labeling, and royalty routing. Developments in 2025–2026—including licensing settlements, GEMA’s litigation against Suno, Spotify’s licensed remix initiative, Music IP Holdings’ patent framework, and investment by major entertainment companies in generative-AI infrastructure—show that governance is increasingly embedded in product architecture and capital relationships. The paper concludes that accountable co-creative systems require permissions, provenance, transparent human contribution, and remuneration mechanisms that reconnect generated outputs to the creative labor and catalogs on which they depend.

Current discussions of artificial intelligence in music often focus on AI-generated artifacts and whether machines can compose autonomously. This paper shifts attention from finished outputs to the broader creative workflow, asking how generative AI can participate in audiovisual composition without directly generating the final score. Through two practice-based case studies, this paper examines AI across conceptual development, prompt mediation, material generation, research, evaluation, and reflection. In the first case, conversational AI supported concept and prompt development, while an image-to-video system produced visual material that required substantial human editing before independently composed music was added. The second case examines a Dunhuang-inspired dance film in which AI supported cultural research, theoretical inquiry, and audiovisual ideation, while verification, composition, orchestration, synchronization, and aesthetic judgment remained human-directed. Drawing on human–AI co-creativity research, Vilém Flusser’s theory of technical apparatuses and play, and Michel Chion’s and Nicholas Cook’s theories of audiovisual relations, the paper proposes a shifting-role co-creative workflow. It argues that AI can shape composition before a single note is written, while artistic judgment, cultural responsibility, and final creative accountability remain human.

Plants and humans interact with the world through profoundly different systems. While plants are known to sense their environment, whether they perceive it remains an open scientific question (Ferretti et al., 2025). Interspecial Signal makes audible the electrical life of a living plant, placing that open question in the listener. A pothos plant is connected to Biotron, a consumer-grade controller that converts changes in the plant's bioelectrical conductivity into MIDI data in real time. That data drives bell-like tones, a radiating visual field, and Somax2, an AI improvisation system trained on a corpus of the author’s own voice, which shadows the plant's signal like an invisible vocalist. Audiences are invited to approach, gesture, and listen, and their proximity shapes what they hear and see. In that shared capacity for sensing, the work offers a point of entry into empathy with a form of life that can otherwise feel wholly alien. This paper documents the work's development across a campus pilot, a conference video, and a museum showing, drawing on audience observations and research on plant awareness disparity. The result is a three-way co-creation between plant, algorithm, and audience, rooted in the proposition that care for the nonhuman world might begin somewhere as small as one “sensational” encounter.

Maintained by Omission is an audiovisual composition that explores the tension between habitual human perception and machine-assisted perception. Rooted in Vilém Flusser’s media philosophy, the project approaches artificial intelligence not as a conscious entity, but as an interactive processing mirror within a cybernetic, dialogic feedback loop. Grounded in the premise that human cognition routinely filters out environmental noise to maintain perceptual stability and social balance, keeping us from being overwhelmed by the world's constant, unfiltered saturation, this paper takes the reader on an inward journey through the project's creative process—mirroring how the composition itself turns inward toward the hidden layers of human perception and daily life. Through an interactive dialogue across language models (ChatGPT, Gemini), audio architectures (Max/MSP Somax 2), generative video models (RunwayML), and AI audio generators (Suno), raw human observations—captured via environmental field recordings, violin improvisations, spoken poetry, and historical audio archives—are calculated, shuffled, recombined, and recomposed. By embracing play and chance over strict control, the project shifts the creative process from a discrete 'chess game' toward an embodied, continuous dance with the machine. Ultimately, Maintained by Omission demonstrates how human-machine co-creation can expose the hidden architectures of daily perception, transforming omitted noise into a rich, shared space of intersubjective synthesis.

An instrument decides what part of the world is permitted to become sound. Each transducer, threshold, and sensor mapping isolates some property of a material and discards the remainder, according to choices made by whoever built the device. Material Synthesis Composition (MSC), the method outlined in this paper, anchors the work in those exact choices: I map the behavior of a material substrate onto compositional structure, and a human performer and a machine listener work within those dynamics. MSC grows out of Speculocultural Technopoiesis (ST), my doctoral framework, which examines how Black speculative traditions and sonic ways of knowing can direct the modification of audio tools. Writing as an experimental media artist, I take composition to include selecting materials, modifying instruments and algorithms, fixing the mappings, and conducting the session. The case is Did Sun Ra Fear Cryosleep?, the water movement of the interactive triptych Organic Memory, in which an earthen orb dissolves in a vessel while a hydrophone, an electret microphone, and three water-quality sensors monitor its disintegration. Five abstractions reduce that chemistry to rhythm, register, tempo, voice count, and drone color; against them, Somax2 improvises from a corpus of NASA sonifications. Drawing on studio sessions with a prototype, the paper sets out the transduction, the mappings, and the recorded corpus I withheld.

Embodied Calligraphy integrates guided and improvised movement, Chinese calligraphic animation, sonification, and AI-mediated interaction through computer vision-based motion tracking and real-time audiovisual feedback. The project explores co-creation among three interconnected agencies: the artist-designer, participants, and interactive technologies. Drawing on intersemiotic translation, multimodal social semiotics, and somaesthetics, the research shows how calligraphic strokes are transformed into bodily gestures and audiovisual expressions, while meaning is distributed through collective interaction. Reciprocal exchanges among dancers, participants, and interactive technologies encourage imitation, improvisation, shared authorship, bodily awareness, and self-fashioning. Based on a participatory workshop at the University of California, Riverside, the study explores how programmed interactions and dancers' guidance influence participants' creative choices and collective engagement. Rather than treating technology solely as a facilitating instrument, Embodied Calligraphy considers its active role in shaping artistic outcomes, proposing an interdisciplinary approach connecting performance, computational creativity, and embodied meaning-making.

Ecosystemic music is a framing of musical systems based on natural ecosystems. Just as energy moves through ecosystems, so too does sound and data move through musical systems. By making this connection between musical systems and ecosystems explicit, I examine how the study of ecosystems can inform the creative and technical choices in the design of these systems. The role of AI in ecosystemic works is twofold: first to enable data and sound processing at a larger scale than has been possible, and second to provide a method of decision making that is distinct from human interpretation. These systems fundamentally propose a method of improvising or performing with AI systems. I examine these dynamics through two approaches to developing systems for co-creativity using algorithmic composition and AI. The first approach is the use of algorithmic systems for recording analysis and playback that respond to real-time inputs such as performer audio or listener presence. The second is the use of autoencoders, either in the preparation or performance stage, to facilitate real-time interaction with data sets and a corpus of recordings.