BLA BLAVATAR vs JAAP BLONK – an experimental collaboration between artist Jonathan Reus and sound poet Jaap Blonk exploring voice dataset making as a performance art form – focusing on situational absurdity, improvisation and bespoke scoring systems. In an uncanny duet, Reus and Blonk call and respond to one another in highly dynamic crescendos. Blonk’s real voice plays against, with, within, and without an AI model of his very own voice. The model, a bespoke real-time performance system developed and performed by Reus, far from the perfectly-voiced AI avatars filling the market, seems to creatively “misbehave” as it collapses transients and fragments of vocalization into relentless sound forms.

Bla Blavatar vs Jaap Blonk emphasizes liveness and the organic growth of a perpetually unfinished work, with reminants of all previous performances carrying forward into the next. The dataset grows, the next model is different, the next score is generated – a process of iteration. The collaboration between Reus and Blonk explores one possibility for how dataset creation can be a deeply musical and social act – a way for people to connect through voice and the mediations of datafication. Alongside the aesthetics of vocal fragmentation, collapse and relentless generation, the process of the performance cycle focuses on the relationship between sound poet Jaap Blonk and artist Jonathan Reus, this work emphasizes trust, mentorship, and shared artistic values that differ fundamentally from the universalizing tendencies of commercial AI.

For this collaboration, Reus worked within the infrastructure of the Intelligent Instruments Lab (IIL) in Iceland, and collaborated with the sound art lab PiNA in Koper as part of the EU S+T+ARTS AIR residency framework. In a close partnership with Victor Shepardson of the IIL, they created Tungnaá, an intentionally open, accessible and hackable real-time AI voice instrument. Like the performance itself, Tungnaá is open and incomplete, the software grows and changes together with the performers and the direction of their musical collaboration – accepting unfinishedness as a desirable quality for musical tools and practices.

The performance is an absurdist take on the cultural obsession with generative AI avatars and the goldrush to capitalise on creative automation. Rather than focusing on the uncanny realism of voice clones, the performance foregrounds the physical, creative and mental effort of performing sound poetry precisely according to an AI-friendly score. Together, Reus and Blonk have invented a number of new sound poetry notation systems meant to be both human and machine readable, carefully balancing ideas of vocal expression and representation with machinic listening. These scores, build up from the full scope of the unicode standard of graphic symbols, is inspired by artistic traditions of typographical scores as well as the traditions of phonetically balanced reading scripts in speech research.

During the performance all of Blonk’s vocalisations are recorded and added to a growing public dataset, and will be used as training data for making Bla Blavatar a better public speaker.

A statement issued by the awards jury of the 2025 Prix Ars Electronica awards called the work “an exemplary demonstration of the symbiosis between artistic practice and technological development” where “artists’ perspectives were deeply involved … throughout the entire AI pipeline”. They additionally praised the work for showcasing “how AI technologies can open entirely new artistic possibilities, serving as a tool for extension rather than substitution”, and “how artists can actively and creatively influence the development of AI, not only by using the technology but also by contributing fundamentally to its design and purpose.”

—

Bla Blavatar vs Jaap Blonk is one outcome of the 2023-24 S+T+ARTS AIR residency Dadasets, which uses interdisciplinary arts-led research to reveal the invisible labor of dataset creation and AI training through the perspective of human voices. This performance premiered at IZIS Festival, an annual event and exhibition organised by PiNA, and one of Slovenia’s premier cultural events presenting international forward-thinking new media art and audio/visual performances. This 12th edition of the festival was curated by Irena Borić with the theme JUST CONTROL.

Made possible with support from:

PiNA – Association for Culture and Education

Intelligent Instruments Lab, University of Iceland

Stroom Den Haag

S+T+ARTS AIR, “Making the Invisible Visible”, is funded by European Union call CNECT/2022/3482066 – Art and the digital: Unleashing creativity for European industry, regions, and society under grant agreement LC-01984767. It is part of the S+T+ARTS programme.