INTERVIEW
Gadi Sassoon
Live at the UN and the Question of What AI Is For
Electronic music rarely finds itself inside rooms like the United Nations. When it does, it is usually discussed rather than performed. Gadi Sassoon’s Live at the UN takes that narrative and shifts that balance by placing sound, gesture and human presence directly into a space more often associated with policy, diplomacy and abstract debate.
Recorded at the closing ceremony of the AI For Good Global Summit in Geneva, the performance sits at the centre of Sassoon’s wider Modes of Vibration project, a body of work that moves freely between album releases, live performance, custom built instruments and software systems. Sassoon approaches the technology as material, something to be shaped, questions and pushed back against rather than the endpoint.
At the core of the live work is TEAL, a real time system that responds to musical gestures as they happen. It doesn't autonomously generate the music, it listens, interprets and reacts which allows Sassoon to improvise while the sound, visuals and also the physical elements on the stage respond to what he does. There is an emphasis on amplification over replacement and on giving an extra dimension for performers to work with rather that automating it away.
Across his work, Sassoon consistently returns to the idea of emotional sovereignty and the question of who controls the emotional arc of an experience, especially in a world that is increasingly shaped by predictive technologies. Instead of allowing systems to optimise or standardise feeling, his work argues for maintaining human authorship with uncertainty and presence at the centre of performance.
The following Q&A explores these ideas in depth. Sassoon discusses everything from performing inside an international institution, the realities of building and touring experimental systems, long term collaborations across the arts and science, and also where he personally draws the line between artistic expression and technological assistance.
Recorded at the closing ceremony of the AI For Good Global Summit in Geneva, the performance sits at the centre of Sassoon’s wider Modes of Vibration project, a body of work that moves freely between album releases, live performance, custom built instruments and software systems. Sassoon approaches the technology as material, something to be shaped, questions and pushed back against rather than the endpoint.
At the core of the live work is TEAL, a real time system that responds to musical gestures as they happen. It doesn't autonomously generate the music, it listens, interprets and reacts which allows Sassoon to improvise while the sound, visuals and also the physical elements on the stage respond to what he does. There is an emphasis on amplification over replacement and on giving an extra dimension for performers to work with rather that automating it away.
Across his work, Sassoon consistently returns to the idea of emotional sovereignty and the question of who controls the emotional arc of an experience, especially in a world that is increasingly shaped by predictive technologies. Instead of allowing systems to optimise or standardise feeling, his work argues for maintaining human authorship with uncertainty and presence at the centre of performance.
The following Q&A explores these ideas in depth. Sassoon discusses everything from performing inside an international institution, the realities of building and touring experimental systems, long term collaborations across the arts and science, and also where he personally draws the line between artistic expression and technological assistance.
Live at the UN is a striking setting for an electronic performance. What was that experience like for you, both creatively and personally?
It was truly exceptional. Working with the UN is quite unique, as you might imagine, and quite intimidating at first, because, like most people, I had no experience with international institutions other than reading about them in the news. Fortunately the team at ITU (International Telecommunications Union) who organise the AI For Good Global Summit are extremely dedicated, forward thinking and supportive, and they have made it possible for me and my team to realise a stage production that really pushes the tech forward while remaining focused on artistic intent, which is not an easy feat on a daily stage programme that lasts a good 12 hours for a whole week. All that while helping me navigate a high-profile summit, where you might bump into anyone from will.i.am to the Swiss Minister of Culture. Creatively, I had the privilege of teaming up with Stefano Polli of Sugo Design, a top visual design company in Italy, who helped realise the visual layer of the TEAL system and directed the visuals during the show, but also provided great feedback about the arrangements of Modes of Vibration I created specifically for this show. It’s the kind of production where you have no room for error: 15 minutes on stage, filmed and streamed live, with a stage tech team operating like a Formula 1 pit team! It’s a real privilege to be on stage knowing that the work of so many talented and dedicated professionals is contributing to what you are doing in that moment. I am happy that the performance went so well that we decided to release it as a live album.
TEAL plays a major role in your live work. How would you describe what TEAL does and how it interacts with you during a performance?
TEAL stands for Transmodal Emotional Amplification Listener. Quite a mouthful, I know, but it’s a bit of a play on words based on something Don Was, the current president of Blue Note, said at the summit’s gala dinner: “AI is just another colour in the palette”. I was brainstorming names for my algorithmic Frankenstein and thought aha! TEAL is the colour that listens. The intuition behind TEAL is that AI can be used as something of a “multidimensional amplifier” based on emotions, instead of a predictor or a generator. The blueprint really is the guitar amp, but revised as a system that takes what a human player performs on an instrument, extrapolates meaning from it on the basis of music theory and gestural analysis, and propagates it to sonic, visual and embodied agents. In simpler terms, it listens to what’s played, interprets musical emotions as taught by me, turns them into stories, and creates sounds, visuals and moves objects on that basis. In other words, I can just improvise on piano or violin, and the system will create a whole bunch of things on stage responding in real-time to what I play.
The sculpture-based synth you use is quite unusual. What first led you to build your own experimental instruments instead of relying on traditional hardware?
For about a decade I have been collaborating with mathematicians and physicists to create algorithms that model the sounds of the real world through a series of research projects in a synthesis technique called physical modelling, which can be conceptualised as the sonic equivalent of ray-tracing in video games. When I was commissioned to create a piece for a Transducer festival at EMPAC in NY, we decided to go the other way, basically creating physical objects from a series of simple algorithms and simulations. It was something of a lighthearted challenge, really. What came out of it are the four plates at the core of my sculpture-based synth, which in turn is the basis of the Modes of Vibration album. It was meant to be a one off experiment, but I was very pleasantly surprised by how strongly people reacted to these strange objects that looked like they belonged in art gallery, but would then be played as an instrument. It snowballed from there!
Your project Modes of Vibration exists as an album, a live performance, and a custom software environment. What inspired you to present the work across these three formats?
It’s really is just the surprising continuation of that snowball into something of an avalanche. Ideas around that instrument and the music it inspired kept multiplying, giving rise to performances, installations and bits of code that eventually coalesced into Tetrad, the software instrument developed in partnership with Physical Audio that serves as the digital double of the physical installation but also expands its capabilities far beyond the scope of the real world, based on the recording and notation techniques I developed to integrate the sculpture-based synth into the album. There is something very liberating about breaking free of the traditional album format and allowing oneself to propagate the vision into other mediums and domains. It’s fascinating to see all the things that multi-modal process gives rise to.
Your past albums involved collaborations with physicists and mathematicians. How has that scientific foundation influenced the ideas behind Modes of Vibration?
As I mentioned, it’s the natural continuation of a collaboration process with scientists that has been ongoing for a decade, with collaborations taking place in Edinburgh, Bologna, London, and Geneva. It’s been very prolific, but never straightforward. Establishing a meaningful dialogue between artistic practice and disciplines like engineering or hard science demands sustained, reciprocal effort to build a shared working language, not just a good idea. While art and science often share conceptual affinities, their concrete methods, vocabularies, and priorities can be misaligned once the collaboration moves into a studio or laboratory. What has been most striking is that, when a flow is eventually established (sometimes after years), entirely new aesthetic territories seem to open up. This process has made my music-making more technically rigorous, yet, somewhat counterintuitively, it has also expanded my creative freedom. The emergence of strict technical frameworks has pushed me into uncharted ground; at times it feels like scoring a science-fiction film with the sensibility of a medieval bard. And I love it.
You work across installations, live shows, electronic releases, and software design. How do those different creative worlds feed into each other?
It’s exactly that: a feedback loop. I find it very fertile to push a vision in all these different directions, and then effectively have one inform the other. It’s not really a sequential process, more of a parallel ecosystem. It’s a good way to test out certain creative ideas too: when they don’t translate across fields, it forces you to make them more compelling.
Your collaboration with Reeps One for the Leipzig Ballet received a lot of positive attention. What did that project teach you about bringing electronic sound into a dance and opera context?
It would have been privilege enough to work with artists of that stature in such a prestigious setting as the Leipzig Opera, but the fact that every performance received 10-minute standing ovations made it absolutely overwhelming. And you’re right, we also got a lot of international press, the whole experience was incredible. I am so grateful to Harry (Reeps One) for having involved me in what is undoubtedly one of the high points of my career as both composer and performer. I am still making sense of it in a lot ways, but one key takeaway is that when you are writing music for dance, you really need to learn how dancers and choreographers hear, listen, count and feel, because as a musician, it really isn’t what you might think. Dance has its own language, and it’s very intricate and sophisticated. It’s incredible to perform alongside such high-level dancers, virtuosos of body and movement: my favourite parts were where I allowed myself space to improvise on violin during the central solos, in a deeply emotional dialogue, and the finale, where I dared sing alongside Harry on this dramatic explosion of bass and strings. Another big takeaway has been that there are very receptive audiences out there for a form of musical innovation that bridges tradition daringly. It was humbling to be in the orchestra pit that usually sits 60 members of the Gewandhaus Orchestra with just Harry and I, performing music we wrote in the same place where Bach is buried, a mere few steps away: but people loved it, so we hit on something quite interesting I think.
There’s a lot of conversation around AI in music right now. Where do you personally draw the line between artistic expression and technological assistance?
I am a bit of an extremist in that regard to be honest. I think AI should be an augmentation, never a replacement. I want AI to be an amplifier. Like the electric guitar vis-à-vis the acoustic guitar. Like distortion turned rock into heavy metal. I am interested in real-time processes that amplify human musicianship, emotions and gestures, and that provide new modes of expression. I’m not interested in generating music with text prompts, any music supervisor can tell you how poor a vehicle for musical ideas words really are. The concerns of the creative communities about the unethical deployment of generative AI are very valid and very real: it is decimating entire industries. Maybe that can be stopped, but if the history of sampling is anything to go by, it will take too long for regulators to really catch up to the tech. At the same time, I worry that we are running the risk of reducing a technology with great creative potential to the ethics of services provided by a handful of companies. The ethics debate is real, but at the same time, deeper I dive, the more I am finding ways to create systems that work in creative ways to help me streamline processes that used to be slow and tedious, or even impossible, creating entirely new ways of making music. And I think there is huge promise in that.
Emotional sovereignty is a theme that comes up in your work. What does that concept mean to you in the era of AI-driven tools and systems?
Emotional sovereignty, to me, is about maintaining authorship over the emotional arc of an experience, both as a performer and as an audience member. In an era where AI systems are increasingly designed to predict, nudge, and optimise our emotional states for engagement metrics, I think artists have a responsibility to create spaces where emotion isn't being harvested but expressed. With TEAL, I'm building tools that read emotional qualities in music in real-time, but the crucial distinction is that this inference serves the performer's intent, not a platform's retention goals. The system amplifies my agency rather than replacing it. I'm still the one deciding what to do with that information, how to shape the experience. There's a difference between AI that listens to you and AI that listens for you. The former extracts; the latter extends. I'm interested in the latter, tools that expand what a human performer can perceive and respond to in the moment, without flattening the complexity of what's actually happening in a room full of people. The real danger isn't AI simulating emotion, it's AI standardizing it. Reducing the infinite texture of human feeling to a handful of categories that can be optimized against. My work pushes back against that by treating emotional inference as a creative material, not a control mechanism.
What do you hope listeners will take away from Live at the UN and the wider Modes of Vibration project as it unfolds?
I hope first and foremost to provide a space where the work allows the audience to bring their own emotions into it. I really hope listeners can find an authentic sense of dialogue in that music. And also that they will see that there are ways to use all this new technology in a way that doesn’t take away our humanity, but supports it and amplifies it.
It was truly exceptional. Working with the UN is quite unique, as you might imagine, and quite intimidating at first, because, like most people, I had no experience with international institutions other than reading about them in the news. Fortunately the team at ITU (International Telecommunications Union) who organise the AI For Good Global Summit are extremely dedicated, forward thinking and supportive, and they have made it possible for me and my team to realise a stage production that really pushes the tech forward while remaining focused on artistic intent, which is not an easy feat on a daily stage programme that lasts a good 12 hours for a whole week. All that while helping me navigate a high-profile summit, where you might bump into anyone from will.i.am to the Swiss Minister of Culture. Creatively, I had the privilege of teaming up with Stefano Polli of Sugo Design, a top visual design company in Italy, who helped realise the visual layer of the TEAL system and directed the visuals during the show, but also provided great feedback about the arrangements of Modes of Vibration I created specifically for this show. It’s the kind of production where you have no room for error: 15 minutes on stage, filmed and streamed live, with a stage tech team operating like a Formula 1 pit team! It’s a real privilege to be on stage knowing that the work of so many talented and dedicated professionals is contributing to what you are doing in that moment. I am happy that the performance went so well that we decided to release it as a live album.
TEAL plays a major role in your live work. How would you describe what TEAL does and how it interacts with you during a performance?
TEAL stands for Transmodal Emotional Amplification Listener. Quite a mouthful, I know, but it’s a bit of a play on words based on something Don Was, the current president of Blue Note, said at the summit’s gala dinner: “AI is just another colour in the palette”. I was brainstorming names for my algorithmic Frankenstein and thought aha! TEAL is the colour that listens. The intuition behind TEAL is that AI can be used as something of a “multidimensional amplifier” based on emotions, instead of a predictor or a generator. The blueprint really is the guitar amp, but revised as a system that takes what a human player performs on an instrument, extrapolates meaning from it on the basis of music theory and gestural analysis, and propagates it to sonic, visual and embodied agents. In simpler terms, it listens to what’s played, interprets musical emotions as taught by me, turns them into stories, and creates sounds, visuals and moves objects on that basis. In other words, I can just improvise on piano or violin, and the system will create a whole bunch of things on stage responding in real-time to what I play.
The sculpture-based synth you use is quite unusual. What first led you to build your own experimental instruments instead of relying on traditional hardware?
For about a decade I have been collaborating with mathematicians and physicists to create algorithms that model the sounds of the real world through a series of research projects in a synthesis technique called physical modelling, which can be conceptualised as the sonic equivalent of ray-tracing in video games. When I was commissioned to create a piece for a Transducer festival at EMPAC in NY, we decided to go the other way, basically creating physical objects from a series of simple algorithms and simulations. It was something of a lighthearted challenge, really. What came out of it are the four plates at the core of my sculpture-based synth, which in turn is the basis of the Modes of Vibration album. It was meant to be a one off experiment, but I was very pleasantly surprised by how strongly people reacted to these strange objects that looked like they belonged in art gallery, but would then be played as an instrument. It snowballed from there!
Your project Modes of Vibration exists as an album, a live performance, and a custom software environment. What inspired you to present the work across these three formats?
It’s really is just the surprising continuation of that snowball into something of an avalanche. Ideas around that instrument and the music it inspired kept multiplying, giving rise to performances, installations and bits of code that eventually coalesced into Tetrad, the software instrument developed in partnership with Physical Audio that serves as the digital double of the physical installation but also expands its capabilities far beyond the scope of the real world, based on the recording and notation techniques I developed to integrate the sculpture-based synth into the album. There is something very liberating about breaking free of the traditional album format and allowing oneself to propagate the vision into other mediums and domains. It’s fascinating to see all the things that multi-modal process gives rise to.
Your past albums involved collaborations with physicists and mathematicians. How has that scientific foundation influenced the ideas behind Modes of Vibration?
As I mentioned, it’s the natural continuation of a collaboration process with scientists that has been ongoing for a decade, with collaborations taking place in Edinburgh, Bologna, London, and Geneva. It’s been very prolific, but never straightforward. Establishing a meaningful dialogue between artistic practice and disciplines like engineering or hard science demands sustained, reciprocal effort to build a shared working language, not just a good idea. While art and science often share conceptual affinities, their concrete methods, vocabularies, and priorities can be misaligned once the collaboration moves into a studio or laboratory. What has been most striking is that, when a flow is eventually established (sometimes after years), entirely new aesthetic territories seem to open up. This process has made my music-making more technically rigorous, yet, somewhat counterintuitively, it has also expanded my creative freedom. The emergence of strict technical frameworks has pushed me into uncharted ground; at times it feels like scoring a science-fiction film with the sensibility of a medieval bard. And I love it.
You work across installations, live shows, electronic releases, and software design. How do those different creative worlds feed into each other?
It’s exactly that: a feedback loop. I find it very fertile to push a vision in all these different directions, and then effectively have one inform the other. It’s not really a sequential process, more of a parallel ecosystem. It’s a good way to test out certain creative ideas too: when they don’t translate across fields, it forces you to make them more compelling.
Your collaboration with Reeps One for the Leipzig Ballet received a lot of positive attention. What did that project teach you about bringing electronic sound into a dance and opera context?
It would have been privilege enough to work with artists of that stature in such a prestigious setting as the Leipzig Opera, but the fact that every performance received 10-minute standing ovations made it absolutely overwhelming. And you’re right, we also got a lot of international press, the whole experience was incredible. I am so grateful to Harry (Reeps One) for having involved me in what is undoubtedly one of the high points of my career as both composer and performer. I am still making sense of it in a lot ways, but one key takeaway is that when you are writing music for dance, you really need to learn how dancers and choreographers hear, listen, count and feel, because as a musician, it really isn’t what you might think. Dance has its own language, and it’s very intricate and sophisticated. It’s incredible to perform alongside such high-level dancers, virtuosos of body and movement: my favourite parts were where I allowed myself space to improvise on violin during the central solos, in a deeply emotional dialogue, and the finale, where I dared sing alongside Harry on this dramatic explosion of bass and strings. Another big takeaway has been that there are very receptive audiences out there for a form of musical innovation that bridges tradition daringly. It was humbling to be in the orchestra pit that usually sits 60 members of the Gewandhaus Orchestra with just Harry and I, performing music we wrote in the same place where Bach is buried, a mere few steps away: but people loved it, so we hit on something quite interesting I think.
There’s a lot of conversation around AI in music right now. Where do you personally draw the line between artistic expression and technological assistance?
I am a bit of an extremist in that regard to be honest. I think AI should be an augmentation, never a replacement. I want AI to be an amplifier. Like the electric guitar vis-à-vis the acoustic guitar. Like distortion turned rock into heavy metal. I am interested in real-time processes that amplify human musicianship, emotions and gestures, and that provide new modes of expression. I’m not interested in generating music with text prompts, any music supervisor can tell you how poor a vehicle for musical ideas words really are. The concerns of the creative communities about the unethical deployment of generative AI are very valid and very real: it is decimating entire industries. Maybe that can be stopped, but if the history of sampling is anything to go by, it will take too long for regulators to really catch up to the tech. At the same time, I worry that we are running the risk of reducing a technology with great creative potential to the ethics of services provided by a handful of companies. The ethics debate is real, but at the same time, deeper I dive, the more I am finding ways to create systems that work in creative ways to help me streamline processes that used to be slow and tedious, or even impossible, creating entirely new ways of making music. And I think there is huge promise in that.
Emotional sovereignty is a theme that comes up in your work. What does that concept mean to you in the era of AI-driven tools and systems?
Emotional sovereignty, to me, is about maintaining authorship over the emotional arc of an experience, both as a performer and as an audience member. In an era where AI systems are increasingly designed to predict, nudge, and optimise our emotional states for engagement metrics, I think artists have a responsibility to create spaces where emotion isn't being harvested but expressed. With TEAL, I'm building tools that read emotional qualities in music in real-time, but the crucial distinction is that this inference serves the performer's intent, not a platform's retention goals. The system amplifies my agency rather than replacing it. I'm still the one deciding what to do with that information, how to shape the experience. There's a difference between AI that listens to you and AI that listens for you. The former extracts; the latter extends. I'm interested in the latter, tools that expand what a human performer can perceive and respond to in the moment, without flattening the complexity of what's actually happening in a room full of people. The real danger isn't AI simulating emotion, it's AI standardizing it. Reducing the infinite texture of human feeling to a handful of categories that can be optimized against. My work pushes back against that by treating emotional inference as a creative material, not a control mechanism.
What do you hope listeners will take away from Live at the UN and the wider Modes of Vibration project as it unfolds?
I hope first and foremost to provide a space where the work allows the audience to bring their own emotions into it. I really hope listeners can find an authentic sense of dialogue in that music. And also that they will see that there are ways to use all this new technology in a way that doesn’t take away our humanity, but supports it and amplifies it.
|
|
Live at the UN documents one point in an ongoing body of work, capturing how these ideas translate in real time within an institutional setting not usually associated with experimental electronic performance. You can check out Live at the UN for yourself right now.
|
|